Three in one: evolution of viviparity, coenocytic placenta and polyembryony in cyclostome bryozoans

Abstract

Background
Placentation has evolved multiple times among both chordates and invertebrates. Although they are structurally less complex, invertebrate placentae are much more diverse in their origin, development and position. Aquatic colonial suspension-feeders from the phylum Bryozoa acquired placental analogues multiple times, representing an outstanding example of their structural diversity and evolution. Among them, the clade Cyclostomata is the only one in which placentation is associated with viviparity and polyembryony—a unique combination not present in any other invertebrate group.

Results
The histological and ultrastructural study of the sexual polymorphic zooids (gonozooids) in two cyclostome species, Crisia eburnea and Crisiella producta , revealed embryos embedded in a placental analogue (nutritive tissue) with a unique structure—comprising coenocytes and solitary cells—previously unknown in animals. Coenocytes originate via nuclear multiplication and cytoplasmic growth among the cells surrounding the early embryo. This process also affects cells of the membranous sac, which initially serves as a hydrostatic system but later becomes main part of the placenta. The nutritive tissue is both highly dynamic, permanently rearranging its structure, and highly integrated with its coenocytic ‘elements’ being interconnected via cytoplasmic bridges and various cell contacts. This tissue shows evidence of both nutrient synthesis and transport (bidirectional transcytosis), supporting the enclosed multiple progeny. Growing primary embryo produces secondary embryos (via fission) that develop into larvae; both the secondary embyos and larvae show signs of endocytosis. Interzooidal communication pores are occupied by 1‒2 specialized pore-cells probably involved in the transport of nutrients between zooids.

Conclusions
Cyclostome nutritive tissue is currently the only known example of a coenocytic placental analogue, although syncytial ‘elements’ could potentially be formed in them too. Structurally and functionally (but not developmentally) the nutritive tissue can be compared with the syncytial placental analogues of certain invertebrates and chordates. Evolution of the cyclostome placenta, involving transformation of the hydrostatic apparatus (membranous sac) and change of its function to embryonic nourishment, is an example of exaptation that is rather widespread among matrotrophic bryozoans. We speculate that the acquisition of a highly advanced placenta providing massive nourishment might support the evolution of polyembryony in cyclostomes. In turn, massive and continuous embryonic production led to the evolution of enlarged incubating polymorphic gonozooids hosting multiple progeny.

Full text

Nekliudovaetal. BMC Ecol Evo (2021) 21:54
https://doi.org/10.1186/s12862-021-01775-z
RESEARCH ARTICLE
Three inone: evolution ofviviparity,
coenocytic placenta andpolyembryony
incyclostome bryozoans
U. A. Nekliudova1,2 , T. F. Schwaha1 , O. N. Kotenko2 , D. Gruber3 , N. Cyran3 and A. N. Ostrovsky2,4*
Abstract
Background: Placentation has evolved multiple times among both chordates and invertebrates. Although they are
structurally less complex, invertebrate placentae are much more diverse in their origin, development and position.
Aquatic colonial suspension-feeders from the phylum Bryozoa acquired placental analogues multiple times, repre-
senting an outstanding example of their structural diversity and evolution. Among them, the clade Cyclostomata
is the only one in which placentation is associated with viviparity and polyembryony—a unique combination not
present in any other invertebrate group.
Results: The histological and ultrastructural study of the sexual polymorphic zooids (gonozooids) in two cyclostome
species, Crisia eburnea and Crisiella producta, revealed embryos embedded in a placental analogue (nutritive tissue)
with a unique structure—comprising coenocytes and solitary cells—previously unknown in animals. Coenocytes
originate via nuclear multiplication and cytoplasmic growth among the cells surrounding the early embryo. This pro-
cess also affects cells of the membranous sac, which initially serves as a hydrostatic system but later becomes main
part of the placenta. The nutritive tissue is both highly dynamic, permanently rearranging its structure, and highly
integrated with its coenocytic ‘elements’ being interconnected via cytoplasmic bridges and various cell contacts. This
tissue shows evidence of both nutrient synthesis and transport (bidirectional transcytosis), supporting the enclosed
multiple progeny. Growing primary embryo produces secondary embryos (via fission) that develop into larvae; both
the secondary embyos and larvae show signs of endocytosis. Interzooidal communication pores are occupied by 1‒2
specialized pore-cells probably involved in the transport of nutrients between zooids.
Conclusions: Cyclostome nutritive tissue is currently the only known example of a coenocytic placental analogue,
although syncytial ‘elements’ could potentially be formed in them too. Structurally and functionally (but not devel-
opmentally) the nutritive tissue can be compared with the syncytial placental analogues of certain invertebrates and
chordates. Evolution of the cyclostome placenta, involving transformation of the hydrostatic apparatus (membranous
sac) and change of its function to embryonic nourishment, is an example of exaptation that is rather widespread
among matrotrophic bryozoans. We speculate that the acquisition of a highly advanced placenta providing massive
nourishment might support the evolution of polyembryony in cyclostomes. In turn, massive and continuous embry-
onic production led to the evolution of enlarged incubating polymorphic gonozooids hosting multiple progeny.
© The Author(s) 2021. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the
original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or
other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this
licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco
mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Open Access
BMC Ecology and Evolution
*Correspondence: a.ostrovsky@spbu.ru; andrei.ostrovsky@univie.ac.at
2 Department of Invertebrate Zoology, Faculty of Biology, Saint
Petersburg State University, Universitetskaja nab. 7/9, 199034 Saint
Petersburg, Russia
Full list of author information is available at the end of the article

Page 2 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
Background
Evolution of parental care is a key novelty affecting off-
spring fitness and survival, influencing the life history
mode and, ultimately, governing the evolutionary suc-
cess of the particular group of organisms [1–3]. One of
the most widespread modes of parental care is the reten-
tion of developing progeny inside the parental body. is
provides protection and is accompanied by an exchange
of gases eventually triggering a bidirectional trans-
port of metabolites between them. e final step is the
acquisition of matrotrophy (extraembryonic nutrition,
EEN)—direct extra-vitelline provisioning of nutrients to
the progeny during incubation [4–7]. is mode is wide-
spread among Animalia, proven or inferred in 22 of 34
phyla, and independently originated between at least 140
and 145 times. In general, the matrotrophic adaptations
of invertebrates are anatomically simpler than those of
chordates, but demonstrate a higher positional, struc-
tural and functional diversity [7, 8].
e most complex matrotrophic mode is placentotro-
phy, and its repeated origins in different chordate and
invertebrate clades (altogether in 16 phyla) indicate the
evolutionary effectiveness of this strategy [4, 6, 7, 9]. Its
benefits include saving the energy required for yolk pro-
duction, reducing the number of eggs, as well as creating
and maintaining a comfortable environment for devel-
oping offspring, resulting in their greater fitness [4, 5].
According to Mossman ([10], p. 156), a placenta is “any
intimate apposition or fusion of the fetal organs to the
maternal tissues for physiological exchange”. Whereas
vertebrate placentae are mostly (although not always) of
a similar origin involving sexual ducts and embryonic
envelopes (reviewed in [4, 6, 9, 11–14]), the placental
analogues of invertebrates originated from a plethora
of tissues and organs and sometimes involve embryonic
envelopes as well. Generally, a ‘placental analogue’ is any
local zone of enhanced nutritional transport develop-
ing during incubation, from simple apposition of non-
specialized epithelia to specialized parental–embryonic
tissue/cell complexes that increase the entire secreting
and absorbing surfaces. In many instances, the special-
ized nutritive structures are formed either by the parent
or by the embryo alone, yet forming a bilateral interface
providing a bidirectional transfer of nutrients (reviewed
in [7]; see also [15]).
e phylum Bryozoa, which are widespread sedentary
colonial suspension-feeders, exhibits the largest propor-
tion of placental species among aquatic invertebrates [7,
16]. Among three bryozoan classes, all phylactolaemates,
all extant stenolaemates and many gymnolaemates
exhibit placentation. is evolved independently once in
both the Phylactolaemata and Stenolaemata, and repeat-
edly in Gymnolaemata [7, 17–21]. eir placental ana-
logues therefore differ in position, structure and origin,
making this phylum an exceptional group for compara-
tive evolutionary studies. Only a few detailed works on
bryozoan placentation are available, all of which focus on
gymnolaemates [15, 22–26].
e order Cyclostomata encompasses living represent-
atives of the class Stenolaemata, one of the most ancient
and successful bryozoan clades that has been docu-
mented since the early Ordovician and remains quite
diverse and abundant in modern marine benthic commu-
nities [27–29]. One of the factors potentially contributing
to stenolaemate success is embryonic incubation, which
has evolved one or possibly more times in their history
in the Paleozoic and Mesozoic (discussed in [16, 30–33]).
Cyclostome incubation chambers (gonozooids) first
appear in the fossil record in the late Triassic and since
then are known in all except one cyclostome family [34,
35].
Reproduction of recent cyclostome bryozoans is
remarkable in combining viviparity (intracoelomic
incubation), placentation and polyembryony [36]. is
unique reproductive pattern may have played an impor-
tant role in their past and current success, making this
group a promising model system for studying the evo-
lution of complex reproductive traits and their signifi-
cance. Our understanding of the reproductive biology of
Cyclostomata, however, is hampered by a severe lack of
data (stressed by [36]). Indeed, apart from the research
on sperm morphology and ultrastructure [37–39], only
five early publications focused on cyclostome oogenesis
and embryonic incubation [40–44]. Several other papers
include additional information [31, 45–54], but these
data are rather fragmentary. Although the first descrip-
tions of cyclostome embryos and larvae were published
more than 150years ago [55–57], only a few papers have
dealt with the embryonic development, larval structure
and metamorphosis of Cyclostomata since then [31, 58–
65]. e most detailed piece of work including all aspects
of cyclostome reproduction from spermiogenesis to lar-
val structure is a monograph by Borg [44], which still
remains the main source of information on this topic. All
aforementioned studies were based on histological tech-
niques that, considering the small size of most cyclos-
tomes, led to many gaps in our knowledge as well as
incorrect interpretations of the morphology and dubious
Keywords: Viviparity, Placenta, Coenocyte, Polyembryony, Evolution, Colonial invertebrates

Page 3 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
or contradictory statements. Only two ultrastructural
studies of cyclostome oogenesis and embryogenesis
(non-published diploma work by Dolinina [66]), and lar-
val microanatomy [67] are known. Recently, Nielsen with
co-authors [68] studied microanatomy of the cyclostome
larva and ancestrula using immunochemistry and confo-
cal laser microscopy.
One of the most intriguing aspects of cyclostome
sexual reproduction is the evolutionary transformation
of the feeding module (autozooid) to the enlarged non-
feeding polymorph (gonozooid), which incubates numer-
ous embryos and larvae (Figs.1, 2). Except for one known
example, cyclostomes are colonial hermaphrodites with
either zooidal gonochory or hermaphroditism and pro-
nounced sexual zooidal polymorphism [16, 20, 36, 69].
Sperm is produced in autozooids, whereas female zooids
(either autozooids or autozooidal polymorphs possess-
ing a polypide—a tentacle crown associated with a gut)
form an ovary, degenerate their polypide and become
gonozooids. ey greatly exceed autozooids in size
and have a modified morphology and anatomy, some-
times merging to form a common incubation chamber.
Embryonic development involves polyembryony when
a single zygote develops to the primary embryo, which
produces numerous (up to 115, [40]) secondary embryos
by fission. As a result, each gonozooid produces multiple,
Fig. 1 General view of mature colonies and their details (stereomicroscope). a, b, insert, Crisia eburnea. a Mature colony having six gonozooids,
mostly with embryos (arrowheads) and one empty (asterisk); b Ramifying branch with gonozooid (arrow: ooeciopore); insert, distal part of
gonozooid showing flattened ooeciopore (arrow). c, d, insert, Crisiella producta. c Mature colony with at least seven gonozooids (arrowheads). d
Distal tip of the branch with gonozooid (arrow: ooeciopore); insert, Distal part of gonozooid showing flattened ooeciopore (arrow)

Page 4 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
genetically identical larvae [70]. Extraembryonic nourish-
ment is provided by the nutritive tissue—a placental ana-
logue of uncertain structure and origin. Apart from the
aforementioned, many key aspects of gonozooid develop-
ment and functioning as well as embryogenesis remain
unstudied. Among them are the source and developmen-
tal stages of the nutritive tissue, nutrient transport from
the autozooids to the gonozooid and from the placenta
to the embryos, the formation of the primary embryo,
mechanisms of polyembryony and larval development.
ese major gaps in our knowledge result in a number of
inaccuracies and even mistakes that have been continu-
ously repeated in the textbooks (see. e.g. [36, 71, 72]).
To fill these gaps and shed light on the morpho-func-
tional transformation of the feeding module to the incu-
bation chamber, we conducted detailed ultrastructural
and microanatomical research of gonozooids in two
species from the cyclostome family Crisiidae (Figs.1, 2).
e focus was on the process of embryonic incubation,
the structure and functioning of the nutritive tissue at its
different developmental stages, embryogenesis and poly-
embryony (Fig. 3). For comparative purposes we addi-
tionally studied autozooid anatomy (Fig.4) (see also [44,
73] and review [74]). In a wider context, the current study
compares placentation in different bryozoan lineages
and discusses placental ‘syncytia’ in both invertebrates
and chordates. We also address the hypothetical conse-
quences of placenta evolution such as polyembryony and
polymorphism, and reconstruct hypothetical scenarios
for the evolution of sexual reproduction in Cyclostomata.
Results
Autozooid
Each cyclostome autozooid consists of a tubular cystid—
body wall of ectocyst (consisting of external cuticle and
calcified wall) underlined by epithelial endocyst, and a
polypide comprising a retractile crown of ciliated ten-
tacles (lophophore) and U-shaped gut (Figs.4, 5a‒c, 6a,
7a). e polypide is suspended within the membranous
sac—a free peritoneal wall with basal membrane dividing
the zooidal cavity into an enclosed coelom with a polyp-
ide, and a pseudocoel—space between the cystid wall and
membranous sac (also termed endo- and exosaccal cavi-
ties, respectively). e latter acts as a hydrostatic system:
contraction of its annular muscles decreases its volume
with a corresponding increase in coelomic fluid pressure,
resulting in lophophore protrusion via zooidal orifice
situated in the centre of terminal membrane in crisiids.
Large paired retractor muscles retract the lophophore
(Figs.4a, c, d, 5a, b, 6a).
Gonozooid
Gonozooids in Crisia eburnea and Crisiella producta,
although differing in shape and size (Figs.1, 2), do not
show detectable differences in anatomy and ultrastruc-
ture. In Crisiidae they develop from rather small tubular
autozooidal female polymorphs (with a polypide having
rudimentary gut) to large club-like incubation cham-
bers. is transformation is accompanied by (1) consid-
erable cystid enlargement (inflation and elongation), and
development of a flattened terminal ooeciostome (tube
for larval release terminated with ooeciopore—skeletal
aperture homologous to autozooidal skeletal aperture)
(Figs.1, 2), (2) irreversible degeneration of the polypide,
and (3) expansion and radical cell rearrangement of the
membranous sac, which becomes part of the placental
analogue (compare Figs.3, 4, 5, 6). e fate of the zooidal
orifice (non-skeletal opening through which the tentacles
are protruded) is unknown, but the folded vestibulum
Fig. 2 Details of gonozooids (stereomicroscope). a, b, Crisia eburnea.
a distal tips of the branches with fully-formed (left) and forming
(right) gonozooids (both containing embryos). b Decalcified tip of
the branch with gonozooid showing embryos and larvae visible
through semitransparent zooidal wall. c Decalcified tip of the
branch with gonozooid containing embryos in Crisiella producta.
Ooeciostomes shown by arrows. e, embryos; l, larva; p, polypide of
autozooid

Page 5 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
Fig. 3 Schematic representations of the gonozooid, nutritive structures and embryonic development in crisiid cyclostomes. a Early primary
embryo surrounded by multilayered ‘envelope’ inside of non-altered membranous sac (corresponding to Figs. 4b, c, 7a, b); b Formation of
coenocytic elements inside and outside of the multilayered ‘envelope’; membranous sac wall is partially incorporated into the developing nutritive
tissue (corresponding to Figs. 4d, 7c‒e). c Early nutritive tissue (corresponding to Figs. 4e, 5a, insert) (in a‒c distal part of growing gonozooid is
not depicted). d Primary embryo producing secondary embryos via fission inside growing nutritive tissue (corresponding to Figs. 4f, 5b, insert, c);
developing coenocytic elements of the membranous sac roof are surrounded by the ‘upper cell complex’ (see Fig. 7f). e Fully-formed gonozooid
filled with secondary embryos and young larvae embedded into the nutritive tissue (corresponding to Figs. 4g, H, 6i, 9i, j); dark solitary cells are
visible on the periphery of the nutritive tissue (see Fig. 9g, h). In all images, basal lamina shown by red line. c, coelom; cw, cystid wall; e, embryo; la,
larva; mc, ‘mesothelial’ cells; ms, membranous sac; nt, nutritive tissue; pc, pseudocoel; uc, ‘upper cell complex’

Page 6 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
Fig. 4 Details of autozooids in Crisiella producta (light microscopy). a Longitudinal section of autozooid with functional polypide (atrial sphincter
shown by white arrow, ‘upper cell complex’ by asterisks, attachment organ by arrowheads, and walls of membranous sac by black arrows). b
Longitudinal section of distal tip of autozooid showing wrinkled lower part of vestibulum, ‘upper cell complex’ around atrial sphincter (one of its
muscles shown by arrowhead) and upper part of the tentacle sheath and membranous sac. c Cross-section of autozooid at the level of atrium. d
Cross-section of autozooid at the level of tentacle crown; membranous sac (smaller arrows) surrounds the tentacle sheath (arrowheads); plugged
communication pore in interzooidal wall shown by larger arrow. a, atrium; c, coelom; ca, caecum; ci, tentacle ciliature; cw, external cystid wall; g,
gut; iw, interzooidal wall; m, mouth; ms, membranous sac; pc, pseudocoel; ph, pharynx; py, pylorus; r, rectum; rm, retractor muscles; t, tentacle; ts,
tentacle sheath; uc, ‘upper cell complex’; v, vestibulum

Page 7 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
Fig. 5 Details of oogenesis and early embryogenesis in Crisiella producta (a‒f Colony branches showing autozooids cross-sectioned on different
levels and gonozooids sectioned in their lower part) (light microscopy). a Two autozooids and gonozooid containing ovary with mature
oligolecithal oocyte (arrow). b‒f Consecutive stages of embryonic growth and placental development. b Gonozooid (upper left) with early primary
embryo surrounded by massive multilayered ‘envelope’ presumably originated from follicular cells. c Gonozooid with early primary embryo showing
loose arrangement of blastomeres and surrounded by multilayered ‘envelope’. d Gonozooid with early primary embryo surrounded by multilayered
‘envelope’ transforming to nutritive tissue. e Gonozooid with primary embryo surrounded by placental analogue (nutritive tissue). Non-altered
membranous sac wall is still visible as a thin ‘nuclei-bearing’ line (arrowheads) either adjacent to or somewhat removed from the multilayered
‘envelope’ or nutritive tissue in b‒e; plugged communication pore in interzooidal wall shown by arrow in e. f Gonozooid with late primary embryo
starting fission; membranous sac (arrowheads) is partially free and partially incorporated into nutritive tissue. In all gonozooids the pseudocoel
(space between placental analogue/membranous sac and cystid wall) is filled with a network of large and irregular, darkly-stained ‘mesothelial’ cells.
c, coelom; cw, external cystid wall; dp, degenerating polypide; e, primary embryo; g, gut; iw, interzooidal wall; ms, membranous sac; nt, nutritive
tissue; ov, ovary; pc, pseudocoel; t, tentacle; ts, tentacle sheath; uc, ‘upper cell complex’

Page 8 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
Fig. 6 Details of gonozooid anatomy and early embryogenesis in Crisiella producta (light microscopy). a Longitudinal section of autozooid with
functional polypide (to the left), developing autozooid with a polypide bud (upper right corner) and developing gonozooid in between; gonozooid
contains extended membranous sac enveloping nearly empty coelomic cavity and early embryo surrounded by nutritive tissue on its bottom
(‘upper cell complex’ in pseudocoel shown by arrows; distal, not yet calcified wall of gonozooid is out of image). a and insert, Early primary embryo
surrounded by developing placental analogue (stage corresponding to Fig. 4e). b Gonozooid with late primary embryo and two secondary
embryos surrounded by nutritive tissue including several large cells and partly incorporating membranous sac (large pseudocoelomocyte in
pseudocoel shown by arrow; gonozooid roof not visible); insert, Gonozooid with primary embryo starting fission (stage corresponding to Figs. 4f
and 5c); in both inserts light areas presumably indicate degenerating cells. c Cross-section of gonozooid with primary embryo starting fission;
membranous sac is non-recognizable, being incorporated into nutritive tissue (plugged communication pore in interzooidal wall shown by arrow).
d Cross-section of the branch with gonozooid in its upper part showing large, yet empty, incubation (coelomic) cavity and the ‘upper cell complex’
in the pseudocoel. In all zooids the pseudocoel is filled with a network of large and irregular, darkly-stained ‘mesothelial’ cells; in b, insert and d
membranous sac shown by arrowheads. c, coelom; cw, external cystid wall; e, primary embryo; g, gut; iw, interzooidal wall; ms, membranous sac; nt,
nutritive tissue; p, polypide; pc, pseudocoel; pp, pseudopore; t, tentacle; uc, ‘upper cell complex’

Page 9 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
Fig. 7 Details of gonozooid ultrastructure in Crisiella producta (a‒d, f) and Crisia eburnea (e, g) (TEM). a ‘Mesothelial’ cell connecting epithelium of
the cystid wall and non-altered membranous sac (here and elsewhere basal lamina of the membranous sac shown by arrowheads). b Interzooidal
wall (autozooid to the left) with communication pore plugged by a pore-cell, and ‘mesothelial’ cells connected with cystid epithelium and
membranous sac. c Processes of ‘mesothelial’ cells passing through basal lamina (asterisks) and contacting binucleate cell of membranous sac; d
‘Mesothelial’ cell adjacent to basal lamina of former membranous sac (short process of the nutritive tissue piercing it and contacting ‘mesothelial’
cell indicated by white arrow). e Presumed pseudocoelomocyte in the pseudocoel. f Interzooidal wall between autozooid (to the left) and
gonozooid with communication pore plugged by two pore-cells (one of them contacting ‘mesothelial’ cells). g Wall of gonozooid with two mural
spines (arrows). c, coelom; cu, cuticle; cw, organic matrix of calcified wall; e, early secondary embryo; ep, epidermal lining of cystid walls; g, gut
of autozooid; iw, interzooidal wall; mc, ‘mesothelial’ cell; ms, membranous sac; nt, nutritive tissue; pc, pseudocoel; pg (1 & 2), pore-cells plugging
communication pore; sc, storage cells; uc, ‘upper cell complex’

Page 10 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
persisted although no lumen was detected inside (Fig.10,
insert).
Similar to the autozooid, the cavity of the early develop-
ing gonozooid is divided into the ‘outer’ pseudocoel and
‘inner’ coelom separated by the wall of the membranous
sac (Figs.3, 4, 5, 6, 7, 8). Multiplication of the embryos
via polyembryony and development of the placental ana-
logue (see below) resulted in a strong reduction of both
cavities, which became diminished to slit-like lacunae.
In the pseudocoel of both, auto- and gonozooids, an
irregular network of presumably mesothelial cells (“mes-
enchymatous cells” of Borg [44], p. 229) is present, being
rather dense in some areas (Figs.5a‒f, 6). It is comprised
of numerous irregular-shaped cells with long processes
that contact the cystid wall epithelium and/or the mem-
branous sac as well as each other (Figs.7a‒d, f, g, 8d, h,
i, 11h). Cell size and shape, the electron density of the
cytoplasm (mostly electron-translucent) and organelle
composition varied in these cells, which contained an
elongated nucleus, autophagosomes and a range of inclu-
sions (small vacuoles, lipid droplets, etc.). Some epithelial
cells of the body wall also formed long processes contact-
ing the ‘mesothelial’ cells.
In both autozooids and gonozooids some of these ‘mes-
othelial’ cells are located adjacent to the basal membrane
of the membranous sac. In gonozooids they were more
numerous, usually flattened and partially covering the
membranous sac as an outer lining (Figs.7a‒d, g, 11g, h).
ese cells appear to be rather active, possessing a large
nucleus mostly filled with euchromatin, numerous mito-
chondria, often cisternae of rough endoplasmic reticu-
lum (RER) and various inclusions—autophagosomes,
lipid droplets (black in TEM images), protein platelets
(grey), microvesicles with electron-transparent content,
etc. eir cytoplasm varied from almost electron-trans-
lucent to electron-dense. Noteworthy, some processes of
the ‘mesothelial’ cells passed through the basal lamina of
the membranous sac and contacted cells/tissue below it
(Fig.7c). In turn, the latter also formed short processes
penetrating the basal lamina and contacting adjacent
‘mesothelial’ cells (Fig.7d). In contrast, in autozooids the
‘mesothelial’ cells associated with the membranous sac
were smaller and contained fewer organelles. No connec-
tions with the cells of the membranous sac were detected.
Some large ‘mesothelial’ cells contained few to many
voluminous (2‒3µm in diameter) spherical inclusions,
either homogeneous (Fig. 7c) or with characteristic
‘striated’ appearance (presumably protein platelets
with paracrystalline structure) (Figs. 7b, f, 8g). Such
cells (termed as ‘storage cells’ here and elsewhere) can
be found in any part of the pseudocoel and were more
numerous in gonozooids. eir cytoplasm was usually
electron-dense, with a large active nucleus, numerous
free ribosomes, mitochondria and RER cisternae.
Similar ‘striated’ inclusions were sometimes visible
in the epithelial cells of the cystid wall and were often
detected in the cells of early larvae (Fig. 10d). At the
same time, they were never observed in the cells of the
membranous sac (of both autozooids and gonozooids)
or of the polypide. In addition, some epithelial cells of
the body wall and some ‘mesothelial’ cells contained
large ‘fine-grained’ electron-translucent zones (Fig.8h)
in both auto- and gonozooids. Similar zones were also
detected in some larval cells (Fig.10f).
Large solitary cells of irregular shape (that could
be termed ‘pseudocoelomocytes’) were occasionally
observed in the pseudocoel of both auto- and gono-
zooids (Figs.6b, 7e). eir cytoplasm varied in electron
density, containing large spherical or lobate nucleus and
organelles (RER cisternae, mitochondria, vacuoles, lipid
droplets and other inclusions) that were distinctly more
numerous than in the ‘mesothelial’ cells.
In autozooids and developing gonozooids of both stud-
ied species, large, darkly stained (in histological sections)
round, oval or cubic cells (termed together the ‘upper
cell complex’ here and elsewhere) were detected in the
pseudocoel surrounding the atrial sphincter and the roof
of the membranous sac (including attachment organ)
(Figs.4a‒c, 5a, 6a, d). ese cells either form a prominent
incomplete ‘ring’-mass or few smaller groups, exhibit-
ing an electron-dense cytoplasm, large nucleus, mito-
chondria, numerous free ribosomes and RER cisternae.
A few of them also bear lipid droplets, autophagosomes
and other inclusions, sometimes with a homogeneous
or ‘striated’ content (Fig. 10g). In the upper part of the
developing gonozooids, the ‘upper cell complex’ filled the
pseudocoel above the roof of the membranous sac with
thickened basal lamina (Figs.6a, d, 10g, h). Cells of the
membranous sac exhibited a similar ultrastructure to
the cells of the ‘upper cell complex’ above them. Later
they become hypertrophied and multiplied their nuclei.
e basal lamina in the roof of the membranous sac was
thicker, more electron-dense and deeply folded than else-
where (Fig.10h). Both cells of the membranous sac and
the ‘upper cell complex’ bore processes, some of which
pierced the basal membrane and contacted the cells on
the opposite side (Fig.7c).
Communication pores are represented by a canal in
the interzooidal wall with a three-dimensional lattice/
labyrinth of small calcified bars (spines). is canal is also
filled by 1‒2, presumably specialized epidermal cell(s)
(Figs. 4d, 6c, 7b, f). In comparison with the ordinary
epithelial cells, the pore-cells are larger, contain a large
lobate nucleus and, often, microfilaments (sometimes
numerous). e cytoplasm’s electron density strongly
varied from transparent to very dense. Some pore-cells

Page 11 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
Fig. 8 Early developmental stages of the nutritive tissue surrounding early primary embryo, and microanatomical details of Crisiella producta (a‒g,
i) and Crisia eburnea (h) (TEM). a Early multilayered ‘envelope’ consisting of mononucleated cells with cytoplasm of contrasting electron density
(membranous sac is out of view). b Fully-formed multilayered ‘envelope’ of mono- and binucleate cells with thin membranous sac lining on its
periphery (here and elsewhere basal lamina of the membranous sac shown by arrowheads). c Partly shown embryo surrounded by the nutritive
tissue contacting wall of non-altered membranous sac. d Peripheral part of embryo (left corner) and nutritive tissue surrounded by the free wall of
non-altered membranous sac. e Embryo lined by multi- and mononucleated cells of modified membranous sac becoming a part of the nutritive
tissue. f Mono- and multinucleated cells in the roof of the membranous sac. g Storage cells in the pseudocoel. h ‘Fine-grained’ areas (arrows) in the
‘mesothelial’ and epithelial cells. i Non-altered wall of the membranous sac with adjoined longitudinal muscle and ring muscle (arrow) embedded
in the basal lamina in young gonozooid (large ‘mesothelial’ cell is visible close to longitudinal muscle). c, coelom; cw, external cystid wall; e, primary
embryo; ep, epidermal lining of cystid walls; iw, interzooidal wall; m, longitudinal muscle; mc, ‘mesothelial’ cell; ms, membranous sac; nt, nutritive
tissue; pc, pseudocoel; sc, storage cell

Page 12 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
possessed numerous mitochondria, autophagosomes and
several Golgi complexes together with small peripheral
vacuoles. If a single cell filled the communication pore,
its opposite terminal parts were wider than the central
narrow part (filling the ‘labyrinth’ canal), thus sometimes
giving the pore-cell a ‘dumbbell’ shape (Fig. 7b). When
two cells plugged the communication pore, their cyto-
plasms may drastically differ in electron density (Fig.7f ).
Some pore-cells were in contact with processes of the
‘mesothelial’ cells.
Mural spines (internal outgrowths of the skeletal wall
directed into the zooidal cavity) are often developed in
the gonozooid roof (Figs.7g, 11i). ey vary in size and
have complex (hooked, forked or branched) shape, being
surrounded by the cells of the ‘upper cell complex’. No
muscles were detected in association with the mural
spines in the studied material.
Very thin annular (ring) muscles are embedded in the
basal lamina of the membranous sac in developing gono-
zooids. Longitudinal muscles were also found close to the
membranous sac in the pseudocoel (Fig.8i), sometimes
adjoining its basal lamina. No muscular elements were
present in mature gonozooids.
Oogenesis
Ovaries were found in Crisiella producta only. e ear-
liest oocytes (20 × 16µm in diameter) were detected in
young gonozooids (= autozooidal polymorphs) with the
polypides having eight (as in autozooids) short tentacles
and an underdeveloped gut with triradial pharynx. No
food was present inside the gut. In our material the early
gonozooids always had only one ovary associated with
the basal part of the gut and containing a single oocyte
surrounded by a single layer of flattened and cubic folli-
cular cells. Mature oocytes (25 × 20µm) are oligolecithal
with few yolk granules in the cytoplasm, and have a large
nucleus with prominent nucleolus (Fig.5a). Incidentally,
single oocytes (20μm in diameter) were found inside the
membranous sac, unconnected to the polypide. Sperm
was not detected in the material studied.
Early embryogenesis andformation ofnutritive tissue
We found early primary embryos in Crisiella pro-
ducta. ey were located in the proximal part of gono-
zooid at the bottom of the membranous sac (Figs.5b‒d,
6a, inserts). Early embryogenesis was accompanied by
polypide degeneration, but no residual polypide (‘brown
body”) was detected except for some resorbing retrac-
tor muscles close to the developing embryo. In contrast,
the membranous sac was greatly expanded, enveloping
the coelomic cavity that occupied most of the growing
gonozooid volume. e ‘upper cell complex’ was present
inside the still voluminous pseudocoel around the upper
part of the membranous sac (Fig.6a, d).
We did not observe the initial stages of cleavage. e
early primary embryo found (diameter approximately
30µm) consisted of about ten large round blastomeres
with loose arrangement (Fig.5b). e blastomeres pos-
sessed an electron-dense cytoplasm and large round
nucleus together with multiple free ribosomes, mito-
chondria and RER cisternae (partly visible in Fig.8a, c, e).
Some blastomeres were clearly resorbing.
e early primary embryo is encircled by a massive,
multilayered ‘envelope’ of large flattened cells that, in
turn, are surrounded by the non-modified wall of the
membranous sac (Figs.3a, 5b, c, 8a, b). ese ‘enveloping’
cells were rather variable in size, thickness and electron
density of their cytoplasm and the number of organelles.
Many of them contained numerous mitochondria; some
had autophagosomes. eir nuclei were large, round or
lobate with large nucleoli. Some of these cells were binu-
cleate. e wall of the membranous sac was either free
or adjoined the multilayered ‘envelope’ surface (Figs.3a,
8b‒d).
At a later stage, coenocytes (sensu [75]) form via
nuclear multiplication and cytoplasmic growth on the
periphery as well as inside the multilayered cellular ‘enve-
lope’ (Fig.5d). e number, size, shape and position of
the nuclei indicate a coenocytic, not syncytial, mecha-
nism of formation (Fig.8c‒f). Similar changes affected
some cells of the membranous sac wall. ey enlarged,
became bi- or multinucleated and were incorporated into
the multilayered structure around the embryo (Figs.3b,
8e). e embryo itself becomes a solid round morula
(about 50µm in diameter) of large, round or cubic cells
(Fig.5e).
As the primary embryo grows, the multilayered ‘enve-
lope’ consisting of mostly individual cells and some
coenocytes is gradually transformed into “nutritive tis-
sue” (term of Harmer [42], p. 133) (Figs.3b, C, 5e, 8c,
d). e individual cells are still recognizable because of
their dark-stained/electron-dense cytoplasm. ese cells
become resorbed later on, leaving large ‘empty spaces’
that are visible in histological sections (Fig.6a, inserts, c).
e wall of the membranous sac surrounding the nutri-
tive tissue is still partly free, partly incorporated into the
nascent placental analogue (Figs.5e, f, 6a‒c, inserts, 8c,
e). e upper (distal) area of the membranous sac wall
that enveloped ‘empty’ coelomic space retains its cellular
structure. e entire membranous sac is surrounded by
the basal lamina (Fig.3c).
Eventually, the proximalmost part of the membranous
sac becomes entirely incorporated into the newly estab-
lished nutritive tissue without detectable cell membranes

Page 13 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
inside (Fig.3c). is tissue is rather uniform ultrastruc-
turally, with electron-translucent cytoplasm and numer-
ous mitochondria, free ribosomes, Golgi complexes and
RER cisternae, indicating prominent synthetic activity
(Fig. 8c, d). Vacuoles with various content, autophago-
somes and numerous lipid droplets of various sizes are
spread through the cytoplasm. e numerous large
nuclei, round as well as lobate, were mostly filled with
euchromatin. Small lipid droplets were also recorded in
some peripheral blastomeres of the embryo embedded in
the nutritive tissue (Fig.8c, d).
As embryos multiplied via fission of primary embryo
(see below), the cells in the roof of the membranous
sac became multinuclear with large lobate nuclei, fur-
ther acquiring coenocytic structure (Fig. 8f) (see also
above). Similar changes occurred in the ‘upper cell com-
plex’. Nonetheless, non-altered peritoneal cells make up
most of the membranous sac wall between its roof and
the lowest part incorporated within the nutritive tissue
around the primary embryo (Fig.3d).
Embryonic ssion
e establishment of the nutritive tissue is accompanied
by embryonic growth and initial embryonic fission. For-
mation, growth and development of the numerous sec-
ondary embryos that constrict off the primary embryo
(polyembryony) occur inside the growing nutritive struc-
ture, which functions as a placental analogue (Fig.3d, e).
When starting fission, the late primary embryo has an
irregular shape forming deep slits separating rounded
and oval prospective secondary embryos (Figs.3d, 5f, 6b,
insert, c). Similar to the early primary embryo, neither
cellular layers nor definite zone(s) of cell proliferation
were detected in this growing embryonic ‘mass’ (Figs.5f,
6b, insert, c). Newly formed secondary embryos differ in
shape and size, and usually present with young larvae in
the same mature gonozooid (Figs.2b, c, 3e, 6b, 10a, b).
Secondary embryos andearly larvae
Young secondary embryos were round or oval solid mor-
ulae of various sizes (smallest embryos were 25 × 20 µm
and 30 × 20 in Crisia eburnea and 30µm and 35 × 25 µm
in Crisiella producta) showing no distinct cell layers
(Figs. 6b, insert, c, 9a, 10a, b). eir blastomeres with
large nuclei were similar in shape and size to the blas-
tomeres of the primary embryo. Some showed signs of
endocytosis: coated pits and small vesicles with vari-
ous content were associated with the plasma membrane
(Fig.9b, insert, d). Interestingly, these pits and vesicles
were present in both peripheral and internal blastomeres.
Possible signs of exocytosis (as tiny foldings of plasma
membrane) were incidentally detected on the surface of
the nutritive tissue (Fig.9b, d).
Growth and development of the morulae were fur-
ther accompanied by delamination via establishment
of the ‘inner’ and peripheral cell zones first (Fig.9c, e).
While embryonic cells in both zones are similar, numer-
ous small, electron-translucent vacuoles appear in asso-
ciation with plasma membranes, marking a ‘border’
between ‘inner’ and peripheral embryonic cells facing
each other (Fig. 9e). e function of these vacuoles is
unclear. Moreover, the apical membrane of the periph-
eral embryonic cells exhibits small vesicles and pits, and
the abutted membrane of the nutritive tissue is folded in
some areas indicating both endo- and exocytosis, corre-
spondingly (Figs.9b, d,11h).
At the next stage, secondary embryos became ellipti-
cal, consisting of 1‒2 layers of larger peripheral, mostly
columnar and prismatic cells lined by the smaller and
flattened, internal cells (former ‘inner’ cells) (Fig.10a, c).
In addition, cells of ‘intermediate shapes’ (cubic or oval)
sometimes occur in between the peripheral and ‘inner’
cells. Flattening of the latter resulted in the formation of
the central slit-like cavity, which subsequently became
more voluminous and moved to the future ‘animal’
embryo pole because of the active cell divisions on the
opposite pole. No signs of cell degradation were recorded
in the central part of the embryo. Both peripheral and
‘inner’ cells have electron-dense cytoplasm containing
large mitochondria. roughout the above stages, the
embryos are embedded in the nutritive tissue.
Next, two-layered embryos form a deep invagination
(future adhesive organ) at the ‘vegetal’ pole, reducing the
central cavity (Fig.10a, b), and begin to form a ciliary
corona with cells containing numerous mitochondria and
common Golgi complexes (Figs.9f, 11c, g). e ciliated
cells develop numerous long (up to 5µm) microvilli near
their bases, which are often branched and anastomosed
(Fig. 9f‒h). Abundant pits and microvesicles together
with some pinocytotic channels were visible in this zone,
indicating active endocytosis between the ciliary bases
(Fig.9h). Ciliated areas often bore a flocculent glycocalyx
of varying thickness located around the cilia bases above
the microvilli (Fig.9g). Some non-ciliated peripheral lar-
val cells develop short, irregular microvilli-like projec-
tions as well.
At the same time the thick cuticular cover develops at
the larval ‘apical’ pole (Fig.10a, b, d). e cuticle con-
sists of an electron-translucent, finely-fibrous, thick inner
layer and electron-dense, very thin outer layer (Fig.10d,
e). e long thin processes of the larval cells occasionally
penetrate the inner cuticle layer. Finally, the deep circular
folding of the cuticular ‘cap’ results in the formation of an
apical invagination (Figs.10a, e).
Cells of the early larvae have large spherical nuclei
with a prominent nucleolus and mostly filled with

Page 14 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
Fig. 9 Details of ultrastructure of the secondary embryos and early larvae in Crisiella producta (a‒e, insert) and Crisia eburnea (f‒h) (TEM).
a Cross-section of the early secondary embryo surrounded by nutritive tissue. b Peripheral part of the early secondary embryo surrounded
by nutritive tissue with folded membrane (asterisks); insert, plasma membrane of embryonic cells with forming coated pit (arrowhead) and
microvesicles. c Part of the early secondary embryo starting delamination. d Apical membrane of peripheral embryonic cells showing forming
coated pit (arrowhead) and (supposedly) endocytic microvesicles (arrows); adjacent membrane of the nutritive tissue strongly folded. e Partial view
of the secondary embryo showing central and peripheral zones during delamination (numerous electron-translucent microvesicles are especially
abundant around the ‘inner’ cells). f Periphery of the early ciliated larva surrounded by nutritive tissue with two large round solitary cells (microvilli
around basal parts of cilia are clearly visible). g Surface of early larva with glycocalyx (arrows) above microvilli. h Surface of early larva with numerous
microvilli, ‘endocytotic’ pits and microvesicles (arrows) between cilia bases. Electron-dense lipid droplets are visible in embryonic and larval cells as
well as in the nutritive tissue. c, coelom; e, secondary embryo; la, larva; m, mitochondrion; n, nucleus; nt, nutritive tissue; rer, rough endoplasmatic
reticulum

Page 15 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
Fig. 10 Details of larval and gonozooid structure in Crisia eburnea (a, h) and Crisiella producta (b‒g) (a, b, light microscopy, c‒h, TEM). a, b Oblique
section of gonozooid showing secondary embryos and early larvae embedded into the nutritive tissue (folded cuticular area of the larval apical
pole shown by arrowheads). c Part of bilayered secondary embryo with central cavity, peripheral columnar and flattened ‘inner’ cell lining. d Apical
part of early larva with cuticular ‘cap’ (asterisks) (arrows show granules with ‘striated’ content). e Part of early larva with folded cuticular ‘cap’ (asterisks).
f Elongated cells of the larval adhesive sac with numerous round inclusions and more peripheral cells with ‘fine-grained’ content (arrowheads).
Electron-dense lipid droplets are visible in embryonic and larval cells. g ‘Upper cell complex’ filling space between gonozooid wall (lower left corner)
and the roof of membranous sac (not shown); folded cuticular wall of remained vestibulum shown in insert. h Cross-section of the uppermost part
of folded roof of coenocytic membranous sac surrounded by thick basal lamina (arrowheads) and ‘upper cell complex’. ao, adhesive organ of larva;
b, bilayered secondary embryo; cw, external cystid wall; la, early larva; m, morula-like secondary embryo; nt, nutritive tissue; pc, pseudocoel; rm, roof
of membranous sac; sc, storage cells; uc, ‘upper cell complex’

Page 16 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
Fig. 11 Details of ultrastructure of fully-formed nutritive tissue in Crisia eburnea (a‒c, e‒g, i, inserts) and Crisiella producta (d, h, j) (TEM). a, b,
Contact of two coenocytic ‘elements’ with lobated nuclei and electron-translucent cytoplasm mostly free of organelles and tight (black arrows)
and adherens (white arrows) junctions connecting neighbouring plasma membranes; insert, presumed endocytotic vesicles (arrows) in the
plasma membrane facing the pseudocoel (here and elsewhere the basal lamina of the former membranous sac shown by arrowheads). c Part of
the nutritive tissue with electron-dense cytoplasm and multiple nuclei in between larvae and secondary embryo. d Periphery of the embryo and
adjacent part of the nutritive tissue with electron-dense cytoplasm, two nuclei, multiple autophagosomes and other inclusions and RER; insert, tight
(black arrow) and adherens (white arrow) junctions connecting neighboring plasma membranes. e Cytoplasmic bridge (white arrowhead) between
coenocytic ‘elements’ (tight and adherens junctions indicated by black and white arrows, respectively). f Interdigitations (asterisks) between
coenocytic ‘elements’ (adherens junctions indicated by white arrows) (in insert and e and f, the adherens junctions are 50 nm wide). g, h, Solitary
cells (arrows) with electron-dense cytoplasm incorporated into the nutritive tissue (basal lamina of the membranous sac shown by arrowheads;
presumed endocytotic vesicles in embryonic cells shown by small arrows). i Longitudinal section of the terminal part of gonozooid with folded roof
of modified (coenocytic) membranous sac covered by the ‘upper cell complex’ (also with coenocytic structure) (skeletal spines shown by arrows);
secondary embryo is recognizable in lower part of image. j Coenocytic ‘upper cell complex’ (to the left) separated by basal lamina from coenocytic
membranous sac. e, secondary embryo; ep, epithelial cell; cw, cystid wall; la, larva; m, mitochondrion; mc, ‘mesothelial’ cell; n, nucleus; nt, nutritive
tissue; rer, rough endoplasmatic reticulum; pc, pseudocoel; uc, ‘upper cell complex’

Page 17 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
euchromatin. ey have electron-dense cytoplasm,
often with numerous RER cisternae, free ribosomes
and mitochondria. Many cells also contained specific
‘fine-grained’ areas in their cytoplasm (Fig. 10f) simi-
lar to those recorded in some epithelial and ‘mesothe-
lial’ cells (Fig.8h). Lipid droplets occur throughout the
cytoplasm together with large protein granules having
a ‘striated’ paracrystalline appearance (Fig. 10d) com-
parable (although smaller) to those found in some ‘stor-
age’ cells (Figs.7b, f, 8g). Cells of the forming adhesive
organ contained numerous spherical granules (presum-
ably proteinaceous), homogeneous or ‘striated’, and
smaller oval granules with ‘opalescent’ content (Fig.10f).
e maximum recorded size of developing larvae was
115 × 110 µm and 120 × 100 µm in C. eburnea and
100 × 75µm in C. producta.
Fully‑formed placental analogue
Embryonic multiplication and growth were accompanied
by a corresponding distal expansion of the nutritive tis-
sue, gradually incorporating the rest of the membranous
sac wall, which still retains its basal lamina. e nutri-
tive tissue with embedded embryos and larvae finally
occupies most of the gonozooid internal space, leaving
only a slit-like pseudocoel between the basal lamina of
the membranous sac and the cystid wall (Figs. 3e, 10a,
b, 11i). e fully-formed placental analogue comprised
large coenocytic ‘parts’/‘areas’ separated by a plasma
membrane and interconnected by cytoplasmic bridges
(Figs.3e, 11a, b, e, f). Large solitary cells are also incor-
porated in the analogue (Figs.3e, 11g, h). ese cells are
usually positioned at the periphery close to the basal
membrane and are round or oval with a slightly convo-
luted membrane. ey have a distinctly electron-dense
cytoplasm and large nucleus (sometimes two), numerous
free ribosomes, mitochondria and RER cisternae. Some
also contained lipid droplets and/or protein platelets.
e ultrastructure of the coenocytic ‘parts’ differs.
All possess large, round or lobate nuclei (sometimes in
groups) filled either with euchromatin or with unevenly
distributed heterochromatin (Fig. 11a‒d). Some coeno-
cytes have an electron-translucent cytoplasm containing
few Golgi complexes and RER cisternae, rare lipid drop-
lets and protein platelets and scattered vacuoles (Fig.11a,
b). Scattered large multivesicular-like bodies and
autophagosomes are present, whereas free ribosomes and
mitochondria are not very abundant. Incidentally, signs
of presumed endocytosis were recorded at the coenocyte
periphery below the basal lamina (Fig.11a, insert).
In contrast, other coenocytic areas appear to be much
more ‘active’, having a more electron-dense cytoplasm
with numerous free ribosomes, mitochondria and vari-
ous inclusions, Golgi complexes and distinctly more
developed RER (Fig. 11c, d). Lipid droplets and protein
platelets (when present) are usually larger too, as are
the autophagosomes. Some areas also contain irregular-
shaped RER cisternae of various sizes filled with electron-
dense material, and vacuoles with flocculent material.
Areas with an ‘intermediate’ appearance are also present.
No clear distributional pattern of the coenocytic ‘parts’
with contrasting ultrastructures was evident through the
gonozooid.
Neighbouring coenocytic areas often ‘overlap’ or form
multiple, sometimes rather complex infoldings (inter-
digitations) (Fig. 11f). In both instances, we recorded
tight and adherens junctions (often alternating in a row)
(Fig.11a, b, insert, e, f). Both interdigitations and junc-
tions tend to be more frequent in the terminal (distal)
part of the gonozooid (see also below). e thickness of
the placental analogue varies regardless of its position.
Narrow areas of the nutritive tissue are present either on
its periphery or between neighbouring embryos/larvae,
preventing them from contacting each other (Figs. 3e,
9c) (minimum thickness about 200nm). ickened parts
are widespread and sometimes composed by two (or
more) adjacent coenocytes, often with contrasting ultra-
structure. In both cases, coenocytes formed short pro-
tuberances piercing the basal lamina and contacting the
‘mesothelial’ cells lining it (Fig.7d).
In the fully-formed functioning gonozooid the ‘upper
cell complex’ and the cells of the roof of the membranous
sac together form a thick coenocytic ‘cap’, separated by a
remarkably thick and folded basal lamina into upper and
lower zones (Figs.3e, 7g, 10h, 11i). Each zone comprises
several coenocytic ‘parts’ having an electron-dense, ribo-
some-filled cytoplasm with abundant mitochondria and
RER elements, and numerous round or lobate (presum-
ably, dividing) nuclei (Fig. 11j). ey also contain lipid
droplets, autophagosomes and, sometimes, granules with
‘striated’ contents (in the ‘upper cell complex’). Inclu-
sions (vacuoles, vesicles, etc.) are generally less numerous
in the ‘cap’ compared to the other parts of the placen-
tal analogue. e adjacent coenocytic ‘parts’ are tightly
packed, showing complex interdigitations with adherens
junctions. e remnants of the folded vestibulum are
often present here in both young and mature gonozooids
(Fig.10, insert).
Discussion
Historical background
Although Borg [44] gave detailed analysis of the previous
studies of the cyclostome reproduction, we feel that the
landmark works should be mentioned here to facilitate a
discussion of our major findings. Smitt [56, 57] was the
first to describe and illustrate multiple embryos [as eggs]
and larvae [as ciliary embryos] inside broken gonozooids

Page 18 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
ofCrisia eburnea and Tubulipora liliacea (as T. serpens).
He termed the soft ‘tissue’ enveloping the embryos as a
“mantle”, measured growing embryos and larvae at dif-
ferent stages of their development, and described the
formation of the adhesive organ in the latter. He also
mentioned that “eggs” develop from the “fatty mass”
inside the gonozooid, which he reported as originating
in the same way as autozooids. Hincks ([76], p. 418) fol-
lowed Smitt, stating that gonozooids are enlarged auto-
zooids “modified for … reproductive functions”.
Harmer applied then the new histological technique
and was the first to describe cyclostome breeding sea-
sons and sexual reproduction in detail in the six species
of three families, Crisiidae [40, 45, 46], Lichenoporidae
[41, 47] and Tubuliporidae [42, 48]. He followed gono-
zooid development in these taxa, detected early germ
cells (both male and female) in the early forming zoo-
ids, and reported oocytes associated with a rudimen-
tary polypide in young gonozooids. He also described
the cleavage and development of the “primary embryo”,
whose further growth was accompanied by the forma-
tion of the “finger-shaped processes” and “embryonic fis-
sion” resulting in multiple “secondary embryos” ([40], p.
200). He also briefly described larval anatomy. Harmer
wrote that embryos and developing larvae were embed-
ded in “nucleated protoplasmic reticulum”—the result of
the transformation of the multinucleated “follicle” (which
he thought was formed from the polypide gut). Consid-
ering the massive embryonic multiplication and growth,
he correctly suggested a nutritive function of the “proto-
plasmic network”, comparing it to the placenta of mam-
mals and salps. Harmer [40] was also the first to describe,
measure and illustrate cyclostome sperm as well as sper-
matogonia and spermatocytes. Noteworthy, in explaining
the life-history of Crisia eburnea, he mentioned that in
spring (e.g., in April) sperm was found in the well-devel-
oped colonies without gonozooids that (“are proably in
most cases of the male sex”), whereas he never found
sperm in the colonies with gonozooids [46, p. 145]. Simi-
lar statements on “male” and “female” colonies were later
published by Robertson [43].
Robertson [43] ‘repeated’ the work of Harmer and
examined four crisiid species. is yielded new infor-
mation on cyclostome reproduction, e.g. she described
and illustrated spermatogenic tissue (as testis) as well
as a multilayered cellular ‘envelope’ around the pri-
mary embryo and detected the formation of the tertiary
embryos in one species (a brief review on cyclostome
reproduction was also included in [77]).
Calvet [62] included data on sexual reproduction and
development of four species from four families (Crisii-
dae, Tubuliporidae, Lichenoporidae, Oncousoeciidae)
in his comprehensive monograph on bryozoan anatomy.
e main focus was on cyclostome embryonic develop-
ment (especially formation of the secondary embryos)
and larval structure and their comparison with those in
gymnolaemates.
Waters [49–51] also provided some data on cyclos-
tome reproduction based on histological sections of five
species from the families Plagioeciidae, Horneridae and
Crisinidae, confirming the presence of polyembryony in
them.
Some preliminary data on sexual reproduction in
Cyclostomata were included in the early work of Borg
[52], who also coined the term “gonozoid”, but his ulti-
mate study on cyclostome reproductive anatomy was
published in 1926 [44]. Borg described the history of the
studies on the incubation chambers in Cyclostomata, and
gave the most detailed and precise description of cyclos-
tome sexual reproduction up until that time. Using his-
tological sections, he described its major aspects in 16
species from seven cyclostome families (Crisiidae, Tubu-
liporidae, Plagioeciidae, Lichenoporidae, Frondipori-
dae, Fasciuculiporidae, Horneridae) and made a careful
comparative analysis, also exhaustively collating his own
results with those of earlier authors. e most detailed
descriptions were devoted to Crisia eburnea and Crisiella
producta, the two species that we studied here.
Finally, a monograph on cyclostome incubation cham-
bers was published by Schäfer [31]. Although it included
some anatomical data on sexual reproduction, including
spermato- and embryogenesis as well as larval structure,
this work mostly focused on the morphological diversity
and evolution of skeletal characters.
Gonozooid development andstructure
Polypide andcoenocytic ‘cap’
Gonozooid development in cyclostomes is accompa-
nied by a considerable enlargement of the autozooid or
(in crisiids) autozooidal polymorph related to the estab-
lishment of the ovary (see below). It includes changes in
cystid size and shape, disintegration of the polypide and
associated organs (retractor muscles and funiculus), and
modification of the hydrostatic apparatus and disintegra-
tion of its muscular elements. e free distalmost part
of the autozooidal cystid (peristome) transforms into an
ooeciostome (specialized tube for larval release) with
corresponding changes to the vestibulum [40, 42], whose
folded cuticle is retained ([62], Pl. 10, Fig.15, oval ‘rugose
bodies’ without designation; our data).
Our study showed that the autozooidal polymorph
(presumptive gonozooid) in Crisiidae contains a non-
feeding polypide still capable of protrusion. Based on the
tentacle length and presence of the vestibulum, an atrial
sphincter (although less developed), ligaments (= atta ch-
ment organ), a tentacle sheath and membranous sac,

Page 19 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
Borg ([44], p. 418) wrote that in Crisiella producta the
“fertile polypides … would be very nearly full-grown”
(presumably, comparing their size with that in autozoo-
ids), although their intestine remains underdeveloped,
which he explained by the influence of the female gonad.
Indeed, all such polypides in C. producta we studied had
shortened tentacles (not seen in C. eburnea). Nonethe-
less, whether other crisiids possess a rudimentary or
non-altered tentacle crown in young gonozooids should
be checked.
e presence of annular muscles in the membranous
sac, longitudinal muscles in the pseudocoel (Fig. 8i)
and retractor muscles ([44]; our data) proves the abil-
ity of the polypide of the prospective gonozooid to
expand and retract. is indicates that sperm uptake
from the water column is its main function. e same is
assumed, for instance, for the cheilostome Celleporella
hyalina, which exhibits non-feeding dwarf female zoo-
ids with a rudimentary polypide and developed hydro-
static apparatus [25, 78]. In both cases the sperm itself
should enter the coelom via a supraneural coelomopore
(reviewed in [79]). Although Borg [44] described and
illustrated a so-called “fertile brown body” in the young
gonozooid of Crisiella producta, we were unable to
confirm its existence, leaving this question open for
further investigation. It is possible that the polypide is
totally utilized for the needs of the growing nutritive
tissue and primary embryo that starts to develop about
that time.
In the gonozooid the wall of the membranous sac
becomes part of the nutritive tissue enveloping embryos
and larvae, while its basal lamina persists. Interestingly,
Borg [44] stated that the membranous sac expands to
almost the entire gonozooid “enclosing the whole mass of
embryos, larvae and nutritive tissue”, and its wall keeps
the “ordinary structure” (pp. 425–426). In contrast,
Harmer ([40], p. 216) wrote that the membranous sac (as
“tentacle sheath”) “probably fuses” with the developing
nutritive tissue, and our data are in agreement with this
view.
Harmer ([40], p. 21) also was the first to describe and
illustrate so-called “distal thickening of the tentacle
sheath”, and it was Borg [44], who recognized it as origi-
nating from the distal part of membranous sac wall and
serving as one of the sources of the nutritive tissue via
cell division, transformation and migration. Our data
complement this view. e ‘distal thickening’ of Harmer
and Borg is what we term the coenocytic ‘cap’ consist-
ing of the strongly modified ‘upper cell complex’ and the
thickened roof of the membranous sac separated by the
basal lamina. During ‘cap’ formation the cells of these two
parts obtain numerous nuclei (also detected by Harmer
[40]) and enlarge, thus becoming coenocytes that contact
each other via short cytoplasmic processes piercing the
basal lamina (Fig.7c, d). It seems likely that the “multi-
nucleated masses of protoplasm” and at least some of the
bi- or multinucleated “giant-cells” described and illus-
trated by Harmer ([40], pp. 220, 240) as “derived from
the thickened distal end of the tentacle-sheath [membra-
nous sac]” are coenocytes which originated from the cells
of the membranous sac roof. Borg ([44], p. 343) also saw
these cells.
e mural spines of various shapes that we detected
as being embedded in the ‘upper cell complex’ in the
terminal part of gonozooids (Figs.7g, 11i), were earlier
described in autozooids of crisiids by Weedon and Tay-
lor [80]. Because the attachment organ of the autozooi-
dal polymorph disappears during transformation to the
gonozooid, their probable function is to anchor the coe-
nocytic ‘cap’ when it becomes part of the massive nutri-
tive tissue enveloping multiple embryos and larvae.
e passage through which mature larvae escape from
the gonozooid was not recognizable in our sections, and
we doubt that such a permanent passage exists. Poten-
tially, mature larvae with actively functioning ciliature
could rupture the coenocytic ‘cap’, reaching the ooecios-
tome. Moreover, we were unable to confirm the existence
of a so-called “larval chamber” isolating ready-to-leave
larvae in the distalmost part of the gonozooid [44]. In our
material, this part was filled by a multicomponent coeno-
cytic ‘cap’ (the upper part of the nutritive tissue) showing
ultrastructural evidence of embryonic nourishment and
also possibly isolating the gonozooid from inflowing sea
water during larval release. Several larvae were situated
beneath this ‘cap’, each surrounded by the cytoplasmic
extensions of the nutritive tissue. We suggest that Borg
was unable to see these very thin cytoplasmic ‘walls’ sur-
rounding larvae in the upper part of the gonozooid in
histological sections, and considered tightly appressed
(but separate) cavities containing the mature larvae as a
single “larval chamber”.
Pseudocoel andinterzooidal pores
e presence of non-feeding gonozooids producing
multiple larvae, indicates directional nutrient trans-
port within cyclostome colonies. e question about its
major route, i.e. interzooidal pores and their structure,
remained unresolved. Harmer ([40], p. 213) was the first
to state that “all the zooecia [zooids] are in organic con-
nection by means of the funicular tissue, which passes
through the pores from one zooecium to another, and
from the zooecia to the ovicell [gonozooid].” Borg ([44],
p. 228) followed Harmer, mentioning that the “strands
of mesenchymatous tissue” in the pseudocoel are con-
nected with the similar “irregular, sparse network” of
the neighbouring zooids “by means of processes which

Page 20 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
pass through the interzooidal pores”. is contradicted
another of Borg’s statements that the interzooidal pores
serve for “exchange of coelomic fluid…and nutritive mat-
ter between various zoids” in cyclostomes, i.e. they are
open ([44], p. 202; [81]).
is contradiction can be explained. e TEM-study
of Nielsen and Pedersen [73] showed both open (distal)
and occluded (basal) interzooidal pores in autozooids of
Crisia eburnea. e authors agreed with Borg that inter-
zooidal pores provide open communication between
neighbour zooids, although some of them are closed at
a later stage by either epithelial cells or by a calcareous
plate (the latter based on an illustration by Boardman and
Cheetham [82]). Moreover, Nielsen and Pedersen [73]
found that closed interzooidal pores exhibited numerous
internal spines and were filled by a single “cincture-cell”
containing microfilaments, mitochondria, Golgi appara-
tus and scant RER—perfectly matching our observations.
A similar structure of the closed interzooidal pores was
also described by Carle and Ruppert [83], who studied C.
elongata. In our material the ‘pore cells’ were sometimes
in contact with processes of the ‘mesothelial’ cells of the
pseudocoel (see also below).
SEM-study of the fine-scale skeletal structures in dif-
ferent cyclostome families by Taylor with co-authors
showed that cyclostomes are characterized by spinelets
developing inside the interzooidal pores and partially
occluding them [84–88]. In some tubuliporine cyclos-
tomes these internal spinelets were so numerous and
tightly appressed (see e.g. [89]) that the pores should be
clearly plugged by interspaced epithelial cells. Elsewhere,
the pores in the young interzooidal walls of Crisulipora
occidentalis are open, but later their lumen becomes
gradually obstructed by calcified spines growing centrip-
etally [80]. Comparing these observations with those of
Nielsen and Pedersen [73] and our data, we suggest that
such a narrowing of the pore lumen inevitably resulted
in its occlusion by 1‒2 bordering epithelial cells (sand-
wiched between the pore spinelets). If so, the initial fluid
(and nutrient) exchange between young neighbour zoo-
ids is substituted by the intracellular transport in the
older ones, suggesting the structural and functional mod-
ification of the ‘pore cells’ for a transport function. e
wide distribution of pore spinelets among cyclostomes
(see above) could indicate that a similar situation is typi-
cal for this group, supposedly in both auto- and gonozoo-
ids. us, the interzooidal pore structure in cyclostomes
is reminiscent of that in gymnolaemates rather than phy-
lactolaemates, as Borg [44, 81] thought.
e cyclostome pseudocoel (exosaccal cavity) is a volu-
minous zooidal cavity that, according to the descriptions
of early authors, contains so-called “funicular tissue” [40]
or “strands of mesenchymatous tissue” ([43, 62]; [44], pp.
229–230). Borg [44] also described it as a network of “sol-
itary mesenchymatous cells” with “long, thread-like pro-
cesses”. Nielsen and Pedersen ([73], p. 66) briefly noted
that the “network of ectodermal cells occupies the basal
part of the cystid”. Moreover, in the studied crisiids, both
Calvet [62] and Borg [44] described “leucocytes” and
“amoebocytes”, correspondingly, in both the pseudocoel
and zooidal coelom.
In contrast, our study on crisiids showed that the
‘mesothelial’ cells comprise a prominent element in the
pseudocoel of both autozooids and gonozooids, forming
a three-dimensional network with strongly varying struc-
ture and density. Its possible functions may be mechani-
cal stabilization of the membranous sac position inside
the cystid in autozooids, as well as an exchange of sub-
stances between pseudocoel and nutritive tissue in func-
tioning gonozooids. Indirect evidence for this transport
is the more numerous and physiologically active ‘meso-
thelial’ cells covering the membranous sac in gonozooids
and contacting the nutritive tissue through cytoplasmic
processes piercing the basal lamina. Similar processes are
also formed by the tissue itself (Fig.7c, d, see also above).
e shape and ultrastructure of the cyclostome ‘mesothe-
lial’ cells are reminiscent of the funicular cells in cheilos-
tomes (e.g. [83]). In both cases, they are mostly elongated
or irregular in shape and have an electron-lucent cyto-
plasm including a few lipid droplets and other organelles.
Also, both may display signs of synthetic activity based
on their RER cisternae and mitochondria [25].
e ‘mesothelial’ cells with homogeneous and striated
paracrystalline inclusions we describe here are probably
‘storage’ cells, accumulating nutrient reserves as charac-
teristic protein platelets (also detected in some larvae)
(compare Figs.7b,d, f, 8g and 10d). ey resemble the
nutrient-storage cells with homogeneous protein plate-
lets found in the zooidal cavity of the placental cheilos-
tome Celleporella hyalina ([25], see also [23, 24]). e
frequent occurrence of such cells in the gonozooids of
Crisia eburnea and Crisiella producta could reflect the
necessity to store nutrients, ensuring against possible
instability of nutrient transport from the autozooids.
At least some of the “spherular leukocytes” depicted by
Calvet [62] in C. denticulata could also be such ‘storage’
cells.
In both crisiid species examined, striated and homoge-
neous inclusions were also detected in some cells of the
‘upper cell complex’ in autozooids and gonozooids. Sur-
prisingly, this massive cell structure was overlooked or
neglected in all previous studies. e different ultrastruc-
ture of the constitent cells leads us to suppose that this
‘complex’ includes the ‘storage’, ‘synthetic’ and ‘totipotent’
(stem) cells. e latter could maintain the cell population

Page 21 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
in the pseudocoel, also participating in coenocytic ‘cap’
formation (see above).
Gonozooid functioning
Gametogenesis andfertilization
In Crisiidae both spermatogonia and oogonia (the latter
5.4μm in diameter, [43]) originate from the mesoderm
(“funicular tissue” in [40]) in the growing zones (com-
mon buds) at the tips of branches, further differentiat-
ing to sperm or oocytes in gonochoristic zooids [40, 43,
44, 62, 66]. Male autozooids contain spermatogenic tis-
sue, whereas most of the female germ cells degenerate;
an ovary is established in a few developing autozooidal
polymorphs (presumptive gonozooids) via association of
the oogonium/early oocyte (normally one, but occasion-
ally two per zooid) with the forming polypide bud. Since
the oocyte is associated with the apical (lower) part of the
polypide bud, growth of the latter results in the oocyte
descending to the bottom of the forming zooid. e
oocyte becomes surrounded by the follicle cells of meso-
dermal origin (“egg-follicle” or “primary follicle” of Borg
[44]) and, together with the polypide, is enclosed inside
the membranous sac. Our study in the White Sea did not
sample the onset of reproduction in the local crisiid pop-
ulation because, as with some other bryozoans studied
[90–92], it starts in late spring under ice. Nonetheless,
in Crisiella producta we were able, apart of zooids with
ovary, to detect a few zooids with early oocytes (20μm in
diameter) unconnected to the polypide and presumably
destined to degenerate [44]. We think that such zooids
grow to form normal autozooids. Similarly, in the ctenos-
tome bryozoan Alcyonidium sp., Faulkner [93] described
the origin of the germ cells from totipotent ones in zooi-
dal buds, followed by ovary formation in prospect female
zooids and degeneration in others.
Ovarian structure is simple. One, rarely two, small oli-
golecithal oocytes are surrounded by a single layer of flat-
tened and, sometimes, cubic follicle cells ([44, 66]; our
data). e diameter of the mature oocytes reportedly var-
ies from 15μm [44] to 25 × 20µm (our data) in Crisiella
producta, and from 17.6μm [40] to 21.6μm [43] in Cri-
sia eburnea. Similar to all other incubating bryozoans
[16, 94], fertilization is intraovarian: the male pronucleus
was reportedly found in the mature oocyte in C. producta
by Borg [44] who also stated that, at the next stage, the
ovary is separated from the modified polypide, and the
oocyte (upon reaching its definitive size) starts to cleave
inside the follicle. is triggers quick degeneration of the
polypide, but the membranous sac persists.
Harmer [40] and Robertson [43] assumed that the for-
mation of the cyclostome gonozooid is triggered by the
establishment of a connection between the oocyte and
the polypide bud. We agree with this suggestion, and it
seems that association of the early oogonium/oocyte with
a polypide bud and establishment of the ovary results in
polypide modification and, thus, shift in the developmen-
tal trajectory of the zooidal bud towards an autozooidal
polymorph (prospective gonozooid). Borg ([44], p. 417)
stressed that the establishment of the ovary results in a
“peculiar transformation of the fertile polypide”.
In addition, Harmer [40] suggested, although indirectly,
that failure of normal gonozooid development and func-
tioning could be explained by failed fertilization. Consid-
ering the idea of fertilization being the main reason for
gonozooid establishment, Borg [44] rejected this stress-
ing that it would be inconsistent with the specific posi-
tion of the gonozooid(s) in a colony and contradicts the
limited number of gonozooids despite massive sperm
production [31, 40, 43, 44, 66]. At the same time, sperm
limitation could be the reason why only one gonozooid
per colony is developed in many cyclostomes and many
full-grown colonies lack them entirely (e.g. [95]). Moreo-
ver, experiments have shown that gonozooid develop-
ment begins but fails in the absence of allosperm in the
crisiid Filicrisia geniculata [69] corroborating the sugges-
tions of Harmer [40].
In contrast, experimental work on a cheilostome
bryozoan and one colonial ascidian demonstrated that
sperm limitation is unlikely for aquatic invertebrates
whose eggs are retained within the maternal body [96].
Accordingly, potentially all autozooidal polymorphs with
established ovaries could be fertilized, but not all will
become gonozooids. Borg [44] suggested that in some
colony areas the autozooids might not provide sufficient
“nutritive conditions” for development and functioning
of multiple gonozooids. Further developing the idea of
colony control over gonozooid development and energy
limitation [36, 97], it is possible to assume that forming
gonozooid(s) that established earlier or in some ‘key loci’
(i.e. having most effective position to get nutrients from
autozooids) could possibly affect younger neighbouring
polymorphs, suppressing their development into gono-
zooids. Taking together, this explains patterns and limita-
tions of colony total resource allocation [90] and specific
position of gonozooids, because each functioning gono-
zooid requires the input of multiple feeding zooids [40,
44, 70, 97]. On the other hand, life at great depth, strong
currents and few available substrates could make sperm
limitation important. We conclude that both sperm limi-
tation [98] and colony control may play roles in different
situations, and that it is premature to assign a single fac-
tor to gonozooid development.
Nutritive tissue: origin andfunctions
According to early descriptions, zygote cleavage inside
the female gonad is accompanied by disintegration of the

Page 22 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
“primary” [i.e. ovarian] and formation of the “secondary”
follicle [44]—a multilayered structure of large squamous
cells that Robertson [43] simply described as a “follicle”.
Distinct ‘layers’ of multiple nuclei in the “multinucleated
follicle surrounding the egg” and early primary embryo
were also depicted by Harmer ([40], pp. 199‒200, Pl.
22, Fig.5). According to Borg ([44], pp. 421‒422) some
constituted cells of the ovarian follicle migrate between
blastomeres (denied by Dolinina [66]), whereas others
disintegrate, and a massive “secondary follicle” (our mul-
tilayered ‘envelope’) is formed around the early primary
embryo via continuous cell ‘detachment’ from the ‘inner
[cell] layer of the membranous sac’. is activity decreases
at some point, and progressively fewer cells are added to
the ‘envelope’, whereas few others migrate between blas-
tomeres. Finally, the “secondary follicle” is substituted by
(i.e. transformed into) the “syncytium” with “irregularly
scattered … numerous small nuclei”, i.e. “nutritive tissue”.
Borg ([44], p. 427) confusingly concluded that the mem-
branous “sac gives origin to all nutritive tissue found in
the gonozooid except the trifting portion of it that origi-
nates from the primary follicle (see also below)”.
Our data confirm that the simple ovarian follicle
(Fig.5a) is substituted by a multilayered cellular ‘enve-
lope’ developing around the embryo (Figs. 5b, 8a, b),
although its origin is still uncertain (see also below). As
the primary embryo grows, the ‘envelope’ around it grad-
ually transforms into the conglomerate of mono- and
binuclear cells and coenocytes (Fig. 8c, d). is recalls
Robertson ([43], p. 136), who described multiplica-
tion of the “small follicle cells” close to the early embryo
“accompanied by a diminution in number of the cells of
the concentric layers” and suggested that the former are
derived from the latter as “a stage in their absorption”. Bi-
and multinucleated cells also appeared in the wall of the
membranous sac (Fig.8e), which becomes incorporated
into the ‘envelope’. Finally, both the multicellular ‘enve-
lope’ and its enveloping part of the membranous sac are
replaced by the multinuclear ‘mass’, which lacks distinct
membrane boundaries within. It is potentially possible
that the coenocytes formed from the cells of both the
multilayered ‘envelope’ and the membranous sac could
merge (thus, yielding a true syncytium), whereas some
cells/coenocytes are resorbed, leaving ‘gaps’ visible as
large empty spaces in our preparations (Fig.6a, inserts).
Most likely, this precise stage was named a “vacuolated
follicle” by Harmer [40], who described it as transitional
during its transformation to the “protoplasmic reticulum
[nutritive tissue]”.
e exact sources of the nutritive tissue remain unclear.
In our TEM-images the wall of non-altered membra-
nous sac consisting of the single cell layer was either
adjoined (Fig.8b), partly contacted (Fig.8c) or free from
the fully-formed multilayered cell ‘envelope’ (Fig. 8d).
Accordingly, the latter potentially could be formed via
dedifferentiation of some cells of the membranous sac
followed by their divisions and displacement (see [44,
66]). Later, some such cells of the ‘envelope’ could further
participate in the formation of the coenocytic elements.
Some cells of the membranous sac wall also become
multinucleated and apparently transformed directly to
coenocytes (Fig.8e).
In addition, at least some of the ovarian (follicle) cells
also could potentially undergo dedifferentiation and divi-
sions, helping form the multilayered ‘envelope’ and coe-
nocytes. Either way, the membranous sac wall (initially,
cellular and later coenocytic) is incorporated into the
forming nutritive tissue. Our scenarios therefore cor-
respond to the opinion of Borg ([44], p. 425) that “the
nutritive tissue derived from the primary and secondary
follicles” [the latter originating from the membranous
sac]. Nonetheless, it remains unclear from Borg’s words
how the ovarian (“primary”) follicle could participate in
this process because he repeatedly mentioned its total
disintegration.
us, the early nutritive tissue surrounding the embryo
should originate from coenocytes developing exclusively
from cells of the membranous sac wall, or from these
together with the cells of the ovarian follicle. Since we
could not detect the membrane boundaries inside the
early nutritive tissue, we suggest that its coenocytic ele-
ments fuse, forming a syncytium. Another possibility
involves a substitution of the multilayered ‘envelope’ by
a single coenocyte, a scenario we consider less probable.
In any event, this situation is remarkable because coeno-
cytes are widely distributed among plants [99–101] and
fungi [102], but known in animals only as temporary
structures in the early development of some cnidarians,
arthropods and teleost fishes ([103–106]; discussed in
[75]) and the parasitic plasmodial stage in orthonectids
[107; G. Slyusarev, personal communication] and myxo-
zoans [108]. We should stress that the latter two cases
convergently mirror situation in cyclostome bryozoans
in having a multinucleated (plasmodial) ‘envelope’ that
encloses and nourises multiple multicellular offspring
(embryos and pansporoblasts). Besides, we are not aware
of syncytia originating from coenocytes in any organism.
Another source of nutritive tissue is cells/coenocytes
originating from the roof of the membranous sac (see
above) in the distal part of gonozooid. e relative con-
tribution of the different cell sources to the definitive
placental analogue requires further study. An impor-
tant aspect is the preservation of the basal lamina of the
membranous sac. is indicates structural and develop-
mental continuity between it and the placental analogue,
with a corresponding shift in function from a mechanical

Page 23 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
(hydrostatic) to nutritive one (discussed below). In addi-
tion, the solitary cells (some binucleated) found within
mature nutritive tissue could be the ‘stem’ cells providing
growth and renewal of the placental analogue throughout
gonozooid formation and functioning. At least some of
the “giant-cells” described by Harmer [40] and Robertson
[43] inside the nutritive tissue could be such stem cells.
Borg [44] thought that the ‘giant-cells’ participate in the
formation of this tissue in tubuliporid cyclostomes.
In summary, the composition of the fully-formed
placental analogue is a complex mosaic in Crisiidae.
Its proximal part (enveloping the primary embryo)
develops as either purely coenocytically or coeno-
cytic-syncytially from one or, possibly, two cell sources
(membranous sac and ovarian follicle). e distal part
is coenocytic, and originates from the membranous
sac and the ‘upper cell complex’. e remaining nutri-
tive tissue is structurally represented by coenocytes and
large solitary cells, both presumably originated from
modified cells of the membranous sac wall. Potentially,
some coenocytic ‘elements’ could be syncytial there.
Interestingly, the early nutritive tissue seems to be more
synthetically active than mono- (often with lobate nuclei)
or binucleated cells of the multilayered ‘envelope’. is
probably reflects their different functions: predominantly
nutrient provisioning versus simultaneous provisioning
and formation of the nutritive tissue. Similarly, the con-
trasting ultrastructure of different coenocytic ‘elements’
constituting the fully-formed placenta seems to reflect
differences in their age and/or functional specialization.
Some parts of the nutritive tissue clearly degrade. Embry-
onic growth and signs of endocytosis (e.g. Fig.9b, d, g,
h) unambiguously point to the nourishing function of the
placental analogue, although the ultrastructure of certain
areas shows that they are not involved in active synthesis.
In contrast with the areas with well-developed RER and
numerous free ribosomes and mitochondria, large parts
of the placental analogue have far fewer such organelles.
Both variants of the coenocytic ‘elements’ had multi-
vesicular-like bodies and autophagosomes (sometimes
numerous) of different size and with different content,
indicating active rearrangement/reparation. Moreover,
the arrangement, size, number and shape of the nuclei
indicate that their divisions continue through the coe-
nocytes, pointing to constant growth. e synthetically
‘non-active’ areas are potentially ‘old’ zones ‘recovering’
after an active period of functioning. Another interpre-
tation is that these areas are predominantly involved in
the bi-directional transport of substances between devel-
oping embryos and the nutritive tissue, supported by the
nutrients accumulating in the pseudocoel. Because the
peripheral zone of the nutritive tissue only rarely showed
signs of endocytosis (Fig.11a, insert), we suggest that it
obtains most of the low-molecular precursors of nutri-
ents from the pseudocoel by facilitated diffusion and
active transport. Intercellular transport via appendages
of the ‘mesothelial’ cells piercing the basal membrane
should be also considered. e placenta accepts (and
partly transforms) nutrients, further transporting them
to the developing offspring. e cytoplasmic bridges
between the adjoining coenocytic ‘elements’ could pro-
vide the necessary transport and redistribution of gases/
nutrients within the entire placenta, whereas various cell
contacts could guarantee the isolation of the embryonic
‘chambers/cavities’ for more effective nourishment—a
general trend in the evolution of matrotrophy (see [7]).
Removal of waste products from developing offspring
should not be neglected.
Brief comparison withother Cyclostomata
Among the main differences in the sexual reproduction
between cyclostome families are the number, shape, posi-
tion and development of gonozooids (reviewed in [31, 44,
109]; see also [95, 110]). In most cases each gonozooid is
a modified autozooid, whereas in the family Lichenopori-
dae a common incubation chamber is formed inside
the colony via resorption of internal calcareous walls
and extension of the membranous sac(s), and nutritive
tissue(s) from the fertile zooid(s) to growing incubation
cavity [36, 41, 44]. e few differences from crisiids are
as follows. Hermaphroditic zooids were described in
Lichenoporidae [41]. e number of oocytes associated
with a polypide in a presumptive gonozooid is normally
one, but may reach six, and at least two primary embryos
were detected in one gonozooid, each in a separate fol-
licle [41, 42, 44, 48, 62]. In tubuliporids and lichenoporids
the gonozooids begin their development as autozooids
with a functional polypide that degenerates upon the
onset of embryogenesis. Development and structure of
the nutritive tissue are similar to crisiids, although for
most groups the data is incomplete. e fission of sec-
ondary embryos into two parts giving rise to the ter-
tiary embryos has been recorded in some tubuliporid,
lichenoporid and hornerid cyclostomes [41, 42, 44], a
feature that was reported first by Robertson [43] for one
crisiid species. is phenomenon requires additional
study.
e only documented extant cyclostome family without
gonozooids is Cinctiporidae. Recently, small, although
macrolecithal, oocytes were discovered in the pharyngeal
wall of Cinctipora elegans. Compared to the large-sized
ancestrulae, this suggests considerable embryonic growth
and, thus, placentation. Incubation presumably occurs
inside very large non-modified autozooids that poten-
tially could hold multiple embryos, i.e. support polyem-
bryony as in other cyclostomes [54].

Page 24 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
Embryonic development
Early embryogenesis
Harmer [40] was the first to note strongly modified
cyclostome development. Before his discovery of poly-
embryony, it was thought that each embryo developed
from a separate egg. e first illustrated description of
embryogenesis was by Smitt [57], who depicted early
secondary embryos consisting of identical cells consti-
tuting, however, the peripheral layer and central mass in
Crisia eburnea. Barrois [58] correctly described the early
(primary) embryo as morula and gave a good descrip-
tion of the larval anatomy and morphogenesis, although
he thought that the initial stage in the formation of the
adhesive organ is a gastrulation. is work was followed
by Metschnikoff [59], who described bilayered (second-
ary) embryos as diblastulae (considering the inner layer
as entoderm) and the stage with invaginating adhesive
organ as pseudogastrula (see also [61]).
Harmer ([40], p. 215), who discovered primary and
secondary embryos in crisiids, described the early cleav-
age as asynchronous, yielding a scattered group of blas-
tomeres that later “come together” to form “a small
rounded mass of undifferentiated embryonic cells [i.e.
morula] …without any…definite layers”. e late primary
embryo was confusingly described as either having an
“ectodermic layer… clearly differentiated” or “obscurely
differentiated onto outer and inner cells” (p. 209). Finally,
when embryonic fission begins, the early secondary
embryos were reported to inherit the bilayered structure.
is view (possibly influenced by the earlier embryo-
logical works mentioned above and further confused by
mentioning the “irregular cavity” of the primary embryo,
p. 209) was supported by most researchers [44, 62, 65,
66, 111] and was included in text-books (e.g. [36, 72]).
In fact, the “irregular cavity” in the distal part of the pri-
mary embryo described by Harmer [40] is constituted
by the deep slits separating future secondary embryos
(Table XXIII, Fig. 11). is author neither mentioned
nor depicted any inner cavity inside the primary embryo.
e same holds true for Calvet [62, p. 335], who only
described the “central cavity” in the secondary embryos.
In contrast, Borg [44, p. 422] wrote about “a small elon-
gated cavity…” as being a “part of a greater, irregular cav-
ity in the interior of the primary embryo”.
Our data on embryonic structure and development dif-
fer substantially from these descriptions, since neither
primary nor early secondary embryos exhibited distinct
layers or cavity in the material studied (Fig.6a, insert).
Analysis of Harmer’s [40–42] illustrations showed no
signs of a bilayered structure. e only indication is the
early embryo depicted in [40], Pl. 23, Fig. 8, showing a
darker periphery. We saw similar staining in a few of our
images (e.g. Fig.5e), but the external and internal embry-
onic cells were otherwise identical. Robertson ([43], p.
141) also wrote that “neither in the primary embryo nor
in the buds when first set free, is there any differentiation
into cell layers… No cavity is present” in C. eburnea (also
stated by Nielsen [64]).
In accordance with that, the early primary embryos had
a loose and irregular cell arrangement, and resorption of
some blastomeres was not uncommon during embryo-
genesis in our histological sections. e latter fact could
explain the statements of the early authors about the
“scattered blastomeres”, “free” blastomeres and “the sepa-
ration of blastomeres and the interpolation of the follicle
cells” during the earliest cleavage stages ([40], p. 215; [43],
p. 136; [44], p. 422). is situation requires checking, e.g.
a similar phenomenon was described in some flatworms
with complex eggs [112] and in salps [113]. Formation of
the secondary embryos occurred without formation of
the finger-like processes but rather via detachment of the
rounded/oval ‘lobes’ from the irregular-shaped late pri-
mary embryo (Figs.5f, 6b, insert, c). Both early primary
and early secondary embryos were compact morulae, and
a bilayered structure via delamination and central cavity
appeared later in the secondary embryos. Differences in
descriptions may be explained by rough fixation methods
applied by previous researchers or, possibly, by misinter-
pretation of the images as the blastomeres in the early
primary embryo are often not clearly distinguished.
Extraembryonic nutrition
Harmer ([40], p. 221) precisely indicated that large pri-
mary embryo and multiple larvae could develop from the
small-sized egg only if the “nutrient material” is trans-
ported via pores from zooids with functional polypides.
He wrote that “nutriment is conveyed to the developing
larvae” by means of the “rich protoplasmic reticulum”
and suggested that the secondary embryos inside its large
“vacuoles” consume the “albuminous fluid” by “diffu-
sion”. Interestingly, Borg [44] thought that the secondary
embryos used cells of the ovarian follicle and some cells
of the multilayered ‘envelope’ migrating between blasto-
meres as nutritive material.
Cells of the primary embryo show no signs of endo-
cytosis on our TEM-images. Some of them, however,
contain small lipid droplets in their near-surface cyto-
plasm which are similar to those in the surrounding early
nutritive tissue (Fig.8c, d). is suggests some nutrient
transport from the placental analogue to the developing
embryo via, e.g. facilitated diffusion and active trans-
port. In contrast, small pits and vesicles associated with
the apical plasma membrane in the peripheral cells of the
early secondary embryos indicate endocytosis. Although
no direct signs of exocytosis were detected on the

Page 25 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
adjoining membrane of the nutritive tissue, that mem-
brane was convoluted in some areas, suggesting such an
opportunity (Fig.9b, insert, d).
Early larvae with an extensive microvillous ‘cover’
developing simultaneously with ciliary corona point to
considerable enlargement of the larval surface and, thus,
intense nutrient consumption. Studies on mammalian
gestation have demonstrated the importance of the sur-
face area through which nutrient transfer occurs [4, 114,
115]. Moreover, numerous pinocytotic canals and vesi-
cles between the bases of the microvilli were detected
(Fig.9f‒h), indicating another mechanism of extraembry-
onic nutrition. Both secondary embryos and early larvae
consume nutrients all around, as was previously shown
for most of the gymnolaemates studied ([15, 25, 26], but
see [116]).
Our data on the microanatomy of early larvae gener-
ally correspond to the description by Nielsen ([64], p.
223) based on histological sections of Crisia eburnea
and Crisiella producta (see also [68]). Among other
details, he mentioned cells with numerous “spherical
bodies of unknown significance” encountered in differ-
ent cell layers. We suppose these to be reserve nutrient
granules accumulated by the larvae during matrotrophic
incubation, and visible due to their size and paracrystal-
line structure. Cells of the adhesive organ on our TEM-
images resemble those on the TEM-microphotographs of
the early larvae of C. denticulata made by d’Hondt [67].
Evolution ofplacentation amongBryozoa
In aquatic invertebrates and chordates the multiple shifts
from a broadcasting (free release of both ova and sperm
into the water column resulting in external fertilization)
to a spermcasting strategy (only sperm is released, fol-
lowed by internal fertilization) were a prerequisite for
acquiring parental care in the form of offspring reten-
tion—either preparitive (viviparity) or post-paritive
(brooding) [5, 16, 20, 117]. Both variants were accom-
panied by the independent evolution of matrotrophy, in
particular, placentotrophy [6, 7].
Among modular suspension feeders, placental ana-
logues are known in a few sponges and entoprocts,
some colonial ascidians, many bryozoans and all salps.
Placentotrophy is also suggested in some acroporid
corals (reviewed in [7]). While the most complex pla-
cental analogues are known in Salpidae ([118], see also
below), bryozoans show a much greater positional and
developmental diversity in this respect, with placentas
originating at least 23 times in all three classes ([7, 17];
see also [18]). In contrast to the majority of placental
bryozoans that are brooders (all phylactolaemates and
most gymnolaemates), all Cyclostomata and the small
gymnolaemate family Epistomiidae are viviparous [16,
19]. Phylactolaemata brood their embryos in an inter-
nal brood sac (reviewed in [20, 119, submitted]), as do
some placental ctenostome and cheilostome Gymnol-
aemata (reviewed in [16, 36, 120]). Some ctenostomes
also brood their progeny in the tentacle sheath modi-
fied into a placental analogue (reviewed in [26]). Finally,
most placental cheilostomes incubate their young in skel-
etal brood chambers (ovicells) (reviewed in [16, 19]). In
all these cases, the brood cavity is external with respect
to the body cavity (e.g. [121, 122]), and cleavage starts
after oviposition of a zygote from the maternal coelom
to the brood chamber via the coelomopore (reviewed in
[79]). e placental analogues are formed by modifica-
tion of the epithelium of the body wall—cystid wall, ooe-
cial vesicle or tentacle sheath, whose function is shifted
from mechanical/protective to nourishing. Notably, in all
matrotrophic cheilostome brooders, cells of the placental
analogue are separated from the embryo by the body wall
cuticle, which is permeable for nutrients [15, 16, 23–25,
78].
In contrast, in cyclostomes embryogenesis begins in
the ovary and continues inside nutritive tissue that sub-
stitutes for the coelomic cavity. In the cheilostome family
Epistomiidae a single embryo develops intracoelomically,
and the placental analogue develops from either the folli-
cle (ovarian) or nurse-cells [22, 123, 124]. In cyclostomes
the major source of placenta is cells derived from the
membranous sac (modified peritoneal lining) and, possi-
bly, the ovarian wall ([40, 44]; our data).
e dynamics of placentation is also quite contrast-
ing: in most cheilostomes and some ctenostomes, pla-
cental analogues function without affecting the polypide
(whether ordinary or rudimentary) recycling: they
develop repeatedly for each embryo and degrade after
larval release. In contrast, in ctenostomes nourishing
embryos in the modified tentacle sheath, the polypide
degenerates before or after deposition of a zygote in the
brooding site (tentacle sheath, brooding pouch) [16, 26,
120, 125]. In Epistomiidae, the polypide degenerates
irreversibly in the enlarged/swollen female zooid that
produces a single larva [22]. Similarly, in cyclostome
bryozoans the placental analogue develops once, being
accompanied by polypide degradation and (gono)zooi-
dal enlargement, but the embryos are numerous. Finally,
the structure of the placental analogues in Gymno- and
Phylactolaemata differs drastically from those in Stenol-
aemata, being cellular and coenocytic (and, possibly, syn-
cytial), respectively. eir development therefore differs
as well: cellular multiplication and hypertrophy versus
nuclear multiplication and cytoplasmic growth.
Despite numerous differences, ultrastructural studies
of placentae indicate common cytological mechanisms
in their functioning. In the placental species of all three

Page 26 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
bryozoan classes, Gymnolaemata, Phylactolaemata and
Stenolaemata, mature placental analogues show an active
synthetic apparatus possessing numerous ribosomes
and mitochondria and developed RER ([15, 24–26, 116,
119, submitted]). High transport and secretory activi-
ties are demonstrated by developed Golgi apparatus
and abundant vesicles as well as by the microvillous or
folded appearance of the apical membranes facing the
embryos. Moreover, early larvae show similar trophic
modifications in all cases. eir surface cells bear devel-
oped microvilli or a network of irregular projections and
show signs of endocytosis (pinocytotic canals and small
vesicles below the cell membrane, sometimes coated pits)
([15, 25, 26, 116, 119, submitted]). As a result of this func-
tional convergence, placental analogues having a differ-
ent positional, developmental and morphological basis
in the representatives of phylogenetically distant groups,
all provide bidirectional transport of substances. In this
respect bryozoan placentation is especially remark-
able, being an excellent example of exaptation—features
that acquired functions for which they were not origi-
nally adapted [126]. Various bryozoan brood chambers
evolved as embryonic protective organs that additionally
became the sites of their nourishment. In some cases,
however, embryos are incubated inside structures (ever-
sible tentacle sheath in ctenostomes, membranous sac
in cyclostomes) that normally participate in polypide
protrusion and retraction, but ultimately became the
nutritive organs in both evolution and ontogeny. is
phenomenon can also be described in terms of a ‘sub-
stitution of function’ accompanied by the corresponding
change in structure.
Functional (and, correspondingly, structural) changes
of organs and organ systems in evolution were exhaus-
tively described and classified by a number of authors
([127–131]; reviewed in [132–136]). Using predomi-
nantly examples among vertebrates, they aimed to for-
mulate the major principles of functional evolution. In
cyclostome bryozoans, the hydrostatic function of the
membranous sac changes to the nutritional one in both
ontogenesis and phylogenesis that can be considered first
as an example of substitution, both morphological [128]
and physiological [131, 132, 134]. e substituting organ
(nutritive tissue) functionally and structurally replaces
the substituted one (membranous sac wall). During this
process some elements of the sac resist (basal mem-
brane), some disappear (ring muscles and probably some
epithelial cells) whereas others are modified (the rest
of the epithelial cells). is situation could be generally
compared, for instance, with that in brooding ophiuroids
incubating their offspring inside the bursae which origi-
nal functions in non-brooding relatives are respiration,
excretion and gamete release. In matrotrophic species
the walls of the bursae seemingly provide nourishment
for the developing larvae (reviewed in [7, 137]).
Both these examples could be also considered as an
extension of functions—obtaining of additional (second-
ary) functions [129]. Some bivalve mollusks brood their
larvae in the specific areas of their gills, and incubation is
accompanied by a reduction of water circulation in them
and, sometimes, acquisition of extraembryonic nutrition
(reviewed in [7]). us, the main (primary) function is
temporarily lost during incubation. In cyclostome gono-
zooids the hydrostatic function is lost irreversibly.
In this context, development of cyclostome nutritive
tissue could be also described as a function change [127]
as weakening of the main or strengthening of an addi-
tional function that results in the restructuring of the
organ morphology and a dominance of secondary func-
tion while the initial main function either becomes sec-
ondary or disappears. Change of function is one of the
most common principles of functional evolution includ-
ing such iconic examples as transformation of the insect
ovipositor to a sting and walking limbs into flippers.
Nowadays the principles of the functional substitution
have further developed being widely applied to the pat-
terns of molecular evolution (reviewed in [138–141]).
Evolution ofcyclostome reproductive pattern
Evolution of matrotrophy triggered a shift in resource
allocation during offspring development in both ver-
tebrates and invertebrates: the higher degree of
extraembryonic nutrition correlates with less intense
vitellogenesis (e.g. [19, 142] and references therein). As
to bryozoans, it was Harmer [40] who compared cyclos-
tome reproduction with that in placental mammals, both
having combination of small, yolk-poor eggs and extra-
vitelline nourishment. It was suggested earlier that the
emergence of placentae resulted in a shift from macro- to
oligolecithal oogenesis in both gymnolaemate groups—
cheilo- and ctenostomes [15–17, 19]. In both cases, an
initial step was presumably the shift from the production
of numerous yolk-poor, small oocytes, a zygote-spawning
strategy and non-incubated planktotrophic larva to mac-
rolecithal oogenesis accompanied by the reduction in
oocyte number, evolution of brooding and lecithotrophic
larva. e acquisition of a placenta resulted in a reversal
to oligolecithal oogenesis accompanied by production of
fewer oocytes.
As to cyclostome bryozoans, we propose two hypothet-
ical scenarios. Numerous female germ cells developing
in the budding zone of modern Cyclostomata [43, 44] is
clearly an ancestral feature, but the number of oogonia/
oocytes associated with the polypide-bud is much less
(1‒6), and only 1–2 oocytes develop to an embryo that
possibly is a secondary trait (reviewed in [16]). If an ovary

Page 27 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
initially produced many small oocytes (as in some zygote-
spawning gymnolaemates), the way to reduce their num-
ber could be by an early start of cleavage directly inside
the ovary (among bryozoans also known in epistomiid
cheilostomes [22] as well as in some matrotrophic cnidar-
ians, gastropods, nemerteans, arthropods, echinoderms
[7] and teleost fishes [143]). is would suppress further
divisions of oogonia and oocytic production. e transi-
tion to intraovarian embryonic development (i.e. vivipar-
ity) potentially could meet the problem of gas exchange
and waste removal, i.e. transport of substances from the
parent to embryos and back. is condition could have
given rise to the origin of placentation via transformation
of the membranous sac (and, possibly, follicle). While
the extraembryonic input to the developing progeny was
small, the number of incubating zooids in a colony was
probably large as occurs in gymnolaemates. Embryos
were incubated in non-modified autozooids and released,
depending on nutrient provisioning, as either feeding or
non-feeding larvae. Evolution of substantial placenta-
tion could result in non-feeding larva and yield polyem-
bryony in enlarged gonozooids [19], and a corresponding
reduction in gonozooid number to a few or just one to
fit the colony energy balance (see also below). In Cincti-
poridae gonozooids were lost as autozooidal size drasti-
cally increased. e eggs became macrolecithal although
remained small in this clade [54].
e second scenario involves an initial shift from oligo-
to macrolecithal oogenesis accompanied by a reduced
number of oocytes and acquision of non-feeding larva
(similar to gymnolaemates). Macrolecithal oocytes of
Cinctipora elegans are apparently produced in the phar-
yngeal wall, but this comparison could be relevant since
the ovary is connected to the polypide in other cyclos-
tomes too. Whether oogenesis may shift to macrol-
ecithal mode in some ancient cyclostomes is unknown.
In any case, the subsequent steps could be the same as
in the first scenario: starting from intraovarian cleav-
age and acquisition of placental analogue, resulting in
polyembryony and the evolution of gonozooids. e
origin of placentation, however, ‘returned’ oogenesis to
the oligolecithal mode. In cinctiporids macrolecithal
mode retained, however, in similar to some placental
gymnolaemates [16, 17, 19]. ese hypotheses could be
potentially tested by plotting data on oogonesis in vari-
ous cyclostome lineages against phylogeny, but currently
insufficient data are available for this work.
Is thecyclostome placenta unique?
Coenocyte placental analogues of cyclostome bryozoans
are the structural and functional (but not developmen-
tal) equivalent of the syncytial placentae known only
in a few invertebrate phyla. Among platyhelminthes,
syncytiality is characteristic of the surface epithelium as
well as the epithelium of the excretory and reproductive
ducts of Cestoda (reviewed in [144, 145]). Placentation
has repeatedly evolved in the phylogenetically distant
cestodes via apposition of the developing embryos to the
uterine syncytium, which forms multilayered cytoplas-
mic outgrowths enveloping individual embryos. In all
studied species with obligatory placentation, the uterine
syncytium shows signs of intense protein synthesis and
secretion (reviewed in [146–148] and references therein).
is ‘embedding’ of embryos in nourishing syncytial ele-
ments is strongly reminiscent of the situation in cyclos-
tome bryozoans.
In the monogenean family Gyrodactylidae the external
tegument, the absorptive layer of the intestine and some
parts of reproductive system, including the uterus, have a
syncytial structure [149, 150]. In the matrotrophic gyro-
dactylids, the surface of embryonic cells adjacent to the
uterus wall shows signs of endocytosis [151]. is resem-
bles the cyclostome secondary embryos absorbing nutri-
ents all over their surface during incubation. Besides,
unlike most described placental analogues (see above),
neither the uterine syncytium of gyrodactylids nor large
parts of the cyclostome nutritive syncytium showed
prominent synthetic activity. Rather, their ultrastructure
indicates active nutrient transport (transcytosis) (e.g.
[149, 152]; our data). Finally, the basal lamina of the gyro-
dactylid uterine lining is thought to participate in nutri-
ent and waste transport as a filter for macromolecules
[149, 152], which might also be the case for the basal lam-
ina of the membranous sac in cyclostome gonozooids.
In almost the entirely matrotrophic onychophoran
family Peripatidae, the syncytial placenta is formed
from uterine epithelium at the site where the embryo is
attached to the uterine wall and further spreading over
most of the uterus as the embryo grows. is syncyt-
ium has a microvillous cover on both sides, presumably
acquiring nutrients from the maternal haemocoel and
transporting them to the embryo cavity via the adja-
cent embryo sac wall; this would make them analogues
to the mammalian noninvasive epitheliochorial placenta
[153–155]. e attachment of the embryo to the uterine
wall presumably induces formation of the placental syn-
cytium. is is comparable to the early formation of coe-
nocytic elements by the membranous sac in cyclostome
bryozoans (triggered by the onset of cleavage).
All salps are placental [156, 157]. e early placenta is
of maternal origin, deriving from a knob of follicle cells
[113, 158]. While growing it merges into the syncytium
and fuses with a modified part of the maternal blasto-
zooid epithelium [157, 159]. Similarly, the cyclostome
zygote starts cleavage within the follicle, which possi-
bly participates in forming the placental analogue. e

Page 28 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
mature placenta of salps consists of two syncytial layers
of maternal origin (the cortical one being secondarily
added by embryonic leucocytes). ese are closely con-
nected by interdigitating microvilli [157], resembling the
coenocytic ‘cap’ developing from the ‘upper cell com-
plex’ and the membranous sac in crisiids. As in cyclos-
tome coenocytes, the placental syncytia of salps contain
numerous mitochondria, vesicles with material of vary-
ing electron density and RER cisternae [157]. is indi-
cates both synthetic and transport activity.
Finally, among vertebrates the hybrid fetomater-
nal syncytial structures are characteristic of all stud-
ied ruminants, lagomorphs and marsupial Perameles.
Among carnivores and bats with endotheliochorial pla-
centa, the maternal uterine epithelium is either cellular
or syncytial, being further displaced by fetal syncytio-
trophoblasts (reviewed in [4]). A syncytial placenta of
solely maternal origin is known in skinks of the genus
Mabuya [160, 161]. Importantly, all aforementioned
syncytial placentae are initially cellular, becoming syn-
cytial via cell fusion. In contrast, the nutritive tissue of
cyclostome bryozoans mostly originates via multiplica-
tion of nuclei and cytoplasmic growth, although coeno-
cytic fusion is potentially possible.
Another important difference is that, despite the
functional commonalities (nutrient synthesis and
transport plus gas exchange and waste removal), all
above-listed placentae develop as derivatives of sexual
organs—modified uterine epithelium or ovary. In con-
trast, the bryozoan placental analogues evolved as a
result of functional substitution, mostly involving no
connection with the reproductive system. For instance,
the nutritive tissue of Stenolaemata, with a possible
contribution by follicular cells, is mainly formed anew
at the basis of the zooidal hydrostatic system (the mem-
branous sac). Similarly, in both placental gymnolaema-
tes and phylactolaemates, embryo nutrition is provided
by a hypertrophied epithelium of the maternal body
wall (see above). From this perspective, this is reminis-
cent of the extensive network of excretory ducts asso-
ciated with the uterus of the cestode Skrjabinacanthus
diplocoronatus; this network distributes large quan-
tities of lipids and is apparently involved in supplying
nutrients to developing eggs [162]. A distant analogy
can be drawn also to matrotrophic sponges, whose
developing embryos are enveloped by the “nutritive fol-
licle”, “placental membrane”, “epilarval trophocyte epi-
thelium”, “nurse cells”, “nutrient capsule”, etc. comprised
of modified cells of the mesohyl, choanocytes, pinaco-
cytes or collencytes (reviewed in [7, 163] and references
therein).
e occurrence of syncytial and coenocytic elements
in both vertebrate and invertebrate placentas raises the
question of their advantages. A syncytial/coenocyte
structure could promote an effective exchange of sub-
stances because of the reduced number of membrane
barriers while maintaining structural integrity. Besides,
a syncytium/coenocyte is capable of rapid increase and
has considerable structural plasticity [4]. is enables
regional differentiation, which could be crucial in the
case of multiple and asynchronous embryo develop-
ment. Nevertheless, cellular placentae co-exist with
syncytial ones, both performing successfully, as evident
in various animal groups. To conclude, syncytial/coe-
nocytic placentae evolved independently in different
phyla, and in each case this seems to be an individual
‘solution’ based on either phylogenetic constrains,
‘developmental potential’ of a particular clade or par-
ticular circumstances arising during gestation.
Polyembryony
e phenomenon of polyembryony—the development of
two to several embryos from one egg—should be inter-
preted as the precocious asexual reproduction via frag-
mentation that begins at the early embryonic stage [164].
It should not be considered as budding, which always
occurs from differentiated somatic cells. In a wider sense,
it should be also viewed as a production of larvae (i.e.
fully-formed autonomous, although non-adult, organ-
isms) from non-equal clusters of blastomeres (embryos)
via some regulatory mechanisms.
e prerequisites and consequences of polyembryony
are intriguing evolutionary issues. e origin(s) of intra-
coelomic incubation (viviparity) in cyclostome bryozo-
ans triggered an entire cascade of changes resulting in
the acquisition of extraembryonic nutrition reflecting
the active interaction between embryo and parent [19].
In turn, evolution of substantial matrotrophy might sup-
port the origin of polyembryony which is thought to be
an integral stage in the life cycle and a synapomorphy of
Cyclostomata [36, 88]. is phenomenon is rather rare in
metazoans but is obligatory in at least 18 taxa from six
phyla and has evolved at least 15 times; most of the cases
involve parasites, with the exception of cyclostomes and
armadillos [164–166]. e “unpopularity” of this repro-
ductive mode is unsurprising given its shortcomings,
primarily a reduced genetic diversity of the offspring as
a result of the multiplication of a single genotype [97, 98,
165].
In the crisiids, however, the genetic variation among
larvae from different gonozooids (i.e. potentially hav-
ing different paternity) in the same colony may be suffi-
cient for their survival [97, 167]. Furthermore, genetically
identical larvae of cyclostomes are released sequentially
over a lengthier period (> 2months) [70]. is enables
repeated testing of the same genotype against changing

Page 29 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
conditions and may be evolutionarily equal to sexual
reproduction testing different genotypes against the same
environment simultaneously [167].
e production of numerous descendants from one/
few zygotes in a colony may compensate for limitations
(e.g., phylogenetic) in egg number or the rarity of ferti-
lization events ([98]; see also above). It may also serve as
an alternative outcome of the trade-off between number
and size of progeny. Polyembryony could be advanta-
geous when brooding females are unable to store sperm,
which may be the case in cyclostomes. More importantly,
it potentially enables enlarging offspring number in
response to increased food availability. For both endopar-
asites and sedentary suspension feeders, food abundance
can change unpredictably. If, however, embryonic mul-
tiplication and growth is extended over time, offspring
number would depend not on initial egg quantity but on
the physiological condition of the parental organism dur-
ing embryogenesis [165]. Moreover, placentation com-
bined with polyembryony in cyclostomes may have the
advantage of fast production of numerous non-feeding
larvae that will establish new colonies during favorable
periods. Based on published data on cheilostome bryo-
zoans, the suggestion is that the duration of offspring
development (oogenesis + embryogenesis) is shorter in
placental versus non-placental brooders because embry-
onic development and growth occur simultaneously [16,
17, 19, 23]. is may be especially important in seasonal
(i.e. less stable) environments, allowing faster occupation
of vacant niches. In habitats with stable conditions (warm
climate or deep-water) this would allow the completion
of more reproductive cycles per year. Indeed, both the
diversity and biomass of cyclostomes around New Zea-
land, the Antarctic, Mediterranean and in Arctic seas
can be very impressive (e.g. [54, 95, 168–170]; Kotenko
& Ostrovsky, personal observations). ese explanations
are applicable to both placental gymnolaemates (with
numerous brooding zooids per colony) and cyclostomes
(with one/few gonozooid(s) producing multiple larvae).
Grbic [166] concluded that, in parasitic insects, the
evolution of all polyembryonic lineages included a tran-
sition from ecto- to endoparasitism, and the possibility
to exploit the host’s nutritive environment for embryo
development led to a change in egg size from large yolky
to small. Similarly, the emergence of matrotrophy and
polyembryony in cyclostome bryozoans might be con-
nected to the shift towards oligolecithal oogenesis (see
above). Another group sharing the unique combination
of viviparity, placentation and obligate polyembryony
with cyclostomes is armadillos Dasypus. In the latter, pol-
yembryony is the only way to overcome the limitations in
uterine shape and of a single implantation site, boosting
the number of offspring [165, 171]. As in cyclostomes,
however, polyembryony could not have occurred without
a preexisting effective mechanism for supplying offspring
with nutrients.
In summary, it is clear that massive larval produc-
tion via polyembryony in Cyclostomata could not have
evolved without the evolution of a highly advanced
placenta. Cyclostome nutritive tissue is a highly inte-
grated system able to quickly grow and rearrange/reju-
venate, effectively delivering and distributing nutrients
to the numerous descendants. Extensive nourishment is
directed from many feeding zooids to one or a few gono-
zooids that then serve as colonial ‘incubation organs’.
is pattern of resource allocation contrasts to that in
phylactolaemate and gymnolaemate bryozoans in which
every incubation chamber is associated with an individ-
ual feeding zooid that is often the main source of nour-
ishment. In our opinion, the massive provisioning of a
very restricted number of ‘consumers’ was a crucial nov-
elty that shaped the evolution of polyembryony. We sug-
gest that quick multiplication of continuously growing
offspring met a requirement to provide additional space
for incubation, resulting in the evolution of cyclostome
gonozooids.
Conclusions
Gonozooid establishment in cyclostomes is probably
governed by the general patterns of colony growth and
resource allocation, although sperm limitation could
also play a role in some cases. In Crisiidae transforma-
tion of zooidal bud to autozooidal polymorph is triggered
by the association of the young oocyte(s) to a polypide
bud whereas the autozooidal polymorph with modified
polypide starts its development to a gonozooid after ferti-
lization and intraovarian beginning of cleavage. Gonozo-
oid formation includes a strong enlargement of the cystid
and the degeneration of the polypide and its hydrostatic
system, although the basal lamina of the membranous sac
persists. e placental analogue is initially a multilayered
cellular ‘envelope’ that is ultimately substituted by the
coenocytic nutritive tissue. It has a complex origin, being
formed from the cells of the membranous sac and, pre-
sumably, follicle cells, both becoming coenocytes. Fusion
of the coenocytes into syncytial elements is potentially
possible too. e fully-formed placental analogue of
crisiids has a unique structure comprising a variety of
coenocytic ‘elements’ of contrasting ultrastructure con-
nected by cytoplasmic bridges and various cell contacts,
and presumed ‘stem’ cells. Ultrastructural evidence dem-
onstrates that the nutritive tissue is a highly integrated
and actively rearranging system involved in both the syn-
thesis and transport of nutrients to developing embryos
and young larvae. Coenocytic placental analogues have
never been previously described in animals.

Page 30 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
Embryogenesis involves polyembryony. e primary
embryo is initially a round morula that becomes irregu-
larly lobed during growth. After separation, the lobes
become secondary embryos without cell differentiation
(with later delamination). Formation of the central cav-
ity and organogenesis yields the larva. Both secondary
embryos and larvae absorb nutrients via endocytosis
and, probably, active transport and facilitated diffusion
because early larvae bear microvilli that greatly increase
the absorptive surface area.
Combining polyembryony with placentation yields
many descendants that can potentially adjust their num-
ber to food availability for the maternal colony. Internal
retention of embryos is a prerequisite for the origin of
placental matrotrophy which, in turn, could stimulate the
evolution of polyembryony and gonozooids in Cyclosto-
mata. Moreover, oligolecithal oogenesis in cyclostomes
might be a consequence of acquiring matrotrophy—a
general trend in matrotrophs, both vertebrate and inver-
tebrate. Despite clear limitations, e.g. reduced genetic
diversity of the offspring and betting on an unverified
genotype, polyembryony can be beneficial for cyclos-
tomes. Production of numerous descendants from one
to a few zygotes may compensate for fewer eggs although
reduces the number of fertile zooids. A lengthier and
sequential release of identical larvae enables repeated
testing of the genotype against changing conditions
and thus may be an evolutionary equivalent to sexual
reproduction.
e complex coenocytic placenta of cyclostome bryo-
zoans is comparable with the syncytial placentae of
certain vertebrates and invertebrates. It shows similar
structural and functional adaptations that enhance nutri-
ent synthesis and trophic interactions with embryos,
albeit differing in origin. Known placentae are mostly
formed via modification of reproductive organs, some-
times with a contribution of embryonic cells. Nutritive
tissue of cyclostomes, although possibly partly originat-
ing from the follicle cells, mostly develops from the mem-
branous sac, which initially has a hydrostatic function.
is example of functional substitution, along with cer-
tain other placental bryozoans, can be considered an
exaptation. e coenocytic/syncytial organization of
placentae may have some advantages such as simpli-
fied nutrient distribution (by reducing the number of
membrane barriers), structural plasticity, rapid growth
and integrity along with flexible synthetic and transport
activities.
In the studied crisiids, the interzooidal pores bear
numerous small internal spines and are filled by one–two
specialized ‘pore cells’. is is presumably the basic mode
in fully-grown cyclostome zooids. In the pseudocoel the
irregular network of ‘mesothelial’ cells (more numerous
and physiologically active in gonozooids) is probably
responsible for maintenance of the membranous sac
shape and position, nutrient accumulation and the intra-
cellular exchange between pseudocoel and coelom/nutri-
ent tissue. ere is also evidence of a direct transfer of
low-weight molecules via the basal lamina of the mem-
branous sac, which may serve as a dynamic filter. We also
interpret both ‘pore-cells and ‘mesothelial’ cells to partic-
ipate in the intercellular transport of substances within a
cyclostome zooids and colonies. e ‘upper cell complex’
is a possible source of totipotent cells for maintaining and
restoring the cell population in the pseudocoel and also
participates in the formation of the coenocytic ‘cap’ in
gonozooids.
Methods
Reproductive colonies of Crisia eburnea (Linnaeus, 1758)
[172] and Crisiella producta (Smitt, 1865a) [56] (Figs.1,
2) growing on kelps and red algae were collected during
the ice-free period from 5 to 15m depth by boat dredging
and SCUBA-diving near the Educational and Research
Station “Belomorskaia”, Saint Petersburg State University
(Chupa Inlet, Kandalaksha Bay, White Sea). Sampling
took place in June and August 2015, from June to Octo-
ber in 2016, in June 2017 and in August 2018.
Freshly collected colonies were fixed in 2.5% glutar-
aldehyde (buffered in 0.1 M Na-cacodylate buffer with
10.26% sucrose, pH 7.4) for 3h and subsequently rinsed
three times in the buffer. Postfixation was done in a 1%
solution of osmium tetroxide (OsO4) in the buffer solu-
tion for 1h followed by three rinses in the buffer. Decal-
cification involved several hours in 5% solution of EGTA
in the buffer. After rinsing in the buffer, all colonies were
dehydrated in a graded ethanol series (30–50–70–80–
90–96%) and in acetone-resin mixtures (3:1–1:1–1:3)
and subsequently embedded in epoxy resin (Agar LVR—
Low Viscosity Resin). Resin blocks were sectioned using
a Reichert UltraCut S microtome with Diatome Histo-
Jumbo and Diatome 35 Ultra diamond knives (Diatome,
Bern, Switzerland). Serial semithin (1.0μm) and ultrathin
(60nm) sections were prepared for examination under
light and transmission electron microscope (TEM),
respectively. Semithin sections were stained with Rich-
ardson’s stain; ultrathin ones were placed on copper grids
and contrasted with 2.5% gadolinium triacetate and 3%
lead citrate. Sections were examined using Zeiss Libra
120 transmission electron microscope (Zeiss, Jena, Ger-
many) and photographed with a digital CCD Olympus
Morada G2 (11 MP, in column) camera.
Some living and fixed colonies were photographed
with a digital Leica DFC295 camera attached to a Leica
M205C stereomicroscope. Altogether, we examined

Page 31 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
approximately 50 gonozooids from 30 colonies anatomi-
cally and ultrastructurally.
Acknowledgements
This study was performed using the laboratories and equipment of the Edu-
cational and Research Station “Belomorskaia”, Saint Petersburg State University,
Russia. Light and transmission electron microscopy was performed at the Core
Facility Cell Imaging and Ultrastructure Research, University of Vienna – mem-
ber of Vienna Life-Science Instruments (VLSI), Austria. We thank Dr Y. Kraus,
Moscow State University, Russia, Dr G. Slyusarev and Dr R. Kostyuchenko, both
Saint Petersburg State University, for valuable information and discussion of
some important aspects of the manuscript, and Ms Y. Dunaeva, Zoological
Institute, Russian Academy of Sciences, Saint Petersburg, Russia, for advice
and help with literature, and Dr M. Stachowitsch, University of Vienna, Austria,
for revising an early draft of the manuscript and improving its English. Two
anonymous reviewers kindly helped to improve the manuscript.
Authors’ contributions
OAN designed and coordinated research. UAN and ONK collected and fixed
material. UAN, ONK, TFS, DG and NC conducted practical work and contrib-
uted to data interpretation. UAN and OAN analyzed the data and wrote the
manuscript. All authors have read and approved the manuscript.
Funding
This study was funded by the Austrian Science Fund (stand-alone project
P27933-B29) (field collecting, histological and ultrastructural research on
oogenesis and placentation) and the Russian Science Foundation (grant
18-14-00086) (field collecting, ultrastructural research on polyembryony, data
processing and manuscript preparation).
Availability of data and materials
The datasets analyzed during the current study are available from the cor-
responding author on reasonable request. All data needed are included in the
paper.
Declarations
Ethics approval and consent to participate
Ethical approval and consent to participate were not required for this study.
No permission was required to collect specimens.
Consent for publication
Consent for publication was not required.
Competing interests
The authors declare that they have no competing interests.
Author details
1 Department of Evolutionary Biology, Integrative Zoology, Faculty of Life Sci-
ences, University of Vienna, Althanstr. 14, 1090 Vienna, Austria. 2 Department
of Invertebrate Zoology, Faculty of Biology, Saint Petersburg State University,
Universitetskaja nab. 7/9, 199034 Saint Petersburg, Russia. 3 Core Facility Cell
Imaging and Ultrastructure Research, Faculty of Life Sciences, University
of Vienna, Althanstr. 14, 1090 Vienna, Austria. 4 Department of Palaeontology,
Faculty of Earth Sciences, Geography and Astronomy, University of Vienna,
Althanstr. 14, 1090 Vienna, Austria.
Received: 9 October 2020 Accepted: 8 March 2021
References
1. Clutton-Brock TH. The evolution of parental care. Princeton: Princeton
University Press; 1991.
2. Lodé T. Oviparity or viviparity? That is the question. Reprod Biol.
2012;12:259–64.
3. Royle NJ, Smiseth PT, Kölliker M, editors. The evolution of parental care.
Oxford: Oxford University Press; 2012.
4. Wooding P, Burton G. Comparative placentation: structures, functions
and evolution. Berlin: Springer; 2008.
5. Avise JC. Evolutionary perspectives on pregnancy. New York: Columbia
University Press; 2013.
6. Blackburn DG. Evolution of vertebrate viviparity and specializations
for fetal nutrition: a quantitative and qualitative analysis. J Morphol.
2015;276:961–90.
7. Ostrovsky AN, Lidgard S, Gordon DP, Schwaha T, Genikhovich G,
Ereskovsky AV. Matrotrophy and placentation in invertebrates: a new
paradigm. Biol Rev. 2016;91:673–711.
8. Trumbo ST. Patterns of parental care in invertebrates. In: Royle NJ,
Smiseth PT, Kölliker M, editors. The evolution of parental care. Oxford:
Oxford University Press; 2012. p. 81–100.
9. Blackburn DG. Convergent evolution of viviparity, matrotrophy, and
specializations for fetal nutrition in reptiles and other vertebrates. Am
Zool. 1992;32:313–21.
10. Mossman HW. Comparative morphogenesis of the fetal membranes
and accessory uterine structures. Contrib Embryol Carnegie Inst Wash-
ington. 1937;26:129–246.
11. Blackburn DG, Evans HE, Vitt LJ. The evolution of fetal nutritional adap-
tations. Forts Zool. 1985;30:437–9.
12. Wourms JP, Grove BD, Lombardi J. The maternal-embryonic relationship
in viviparous fishes. Fish Physiol. 1988;11B:1–134.
13. Lombardi J. Postzygotic maternal influences and the maternal-embry-
onic relationship of viviparous fishes. Am Zool. 1996;36:106–15.
14. Blackburn DG. Classification of the reproductive patterns of amniotes.
Herpetol Monogr. 2000;14:371–7.
15. Moosbrugger M, Schwaha T, Walzl MG, Obst M, Ostrovsky AN. The
placental analogue and the pattern of sexual reproduction in the
cheilostome bryozoan Bicellariella ciliata (Gymnolaemata). Front Zool.
2012;9:1–20.
16. Ostrovsky AN. Evolution of sexual reproduction in marine invertebrates:
example of gymnolaemate bryozoans. Dordrecht: Springer; 2013.
17. Ostrovsky AN, Gordon DP, Lidgard S. Independent evolution of matro-
trophy in the major classes of Bryozoa: transitions among reproduc-
tive patterns and their ecological background. Mar Ecol Prog Ser.
2009;378:113–24.
18. Lidgard S, Carter MC, Dick MH, Gordon DP, Ostrovsky AN. Division of
labor and recurrent evolution of polymorphisms in a group of colonial
animals. Evol Ecol. 2012;26:233–57.
19. Ostrovsky AN. From incipient to substantial: evolution of placentotro-
phy in aquatic colonial invertebrates. Evolution. 2013;67:1368–82.
20. Ostrovsky AN. Sexual reproduction in Bryozoa. In: Schwaha TF, editor.
Handbook of zoology. Bryozoa. Berlin: De Gruyter; 2020. p. 101–21.
21. Schwaha T, Ostrovsky AN, Wanninger A. Key novelties in the evolution
of aquatic colonial phylum Bryozoa: evidence from soft-body morphol-
ogy. Biol Rev. 2020;95:696–729.
22. Dyrynda PEJ, King PE. Sexual reproduction in Epistomia bursaria
(Bryozoa, Cheilostomata), an endozooidal brooder without polypide
recycling. J Zool. 1982;198:337–52.
23. Dyrynda PEJ, King PE. Gametogenesis in placental and nonplacental
ovicellate cheilostome Bryozoa. J Zool. 1983;200:471–92.
24. Hughes DJ. Gametogenesis and embryonic brooding in the cheilos-
tome bryozoan Celleporella hyalina. J Zool. 1987;212:691–711.
25. Nekliudova UA, Schwaha TF, Kotenko ON, Gruber D, Cyran N, Ostrovsky
AN. Sexual reproduction of the placental brooder Celleporella hyalina
(Bryozoa, Cheilostomata) in the White Sea. J Morphol. 2019;280:278–99.
26. Schwaha T, Moosbrugger M, Walzl M, Ostrovsky AN. First ultrastructural
evidence of placental nutrition in a ctenostome bryozoan: example of
Amathia verticillata. Zoomorphology. 2019;138:221–32.
27. Boardman RS. General features of the class Stenolaemata. In: Robinson
RA, editor. Treatise on invertebrate paleontology. Bryozoa (Part G,
revised). Boulder: Geological Society of America, Lawrence: University
of Kansas Press; 1983. p. 49–137.
28. Cook P, Bock PE, Gordon DP, Weaver H, editors. Australian Bryozoa. Biol-
ogy, ecology and natural history. Melbourne: CSIRO Publishing; 2018. p.
1–194.
29. Ernst A. Fossil record and evolution of Bryozoa. In: Schwaha TF, editor.
Handbook of zoology. Bryozoa. Berlin: De Gruyter; 2020. p. 11–55.

Page 32 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
30. Taylor PD, Larwood GP. Major evolutionary radiations in the Bryozoa. In:
Taylor PD, Larwood GP, editors. Major evolutionary radiations. Oxford:
Clarendon Press; 1990. p. 209–33.
31. Schäfer P. Brutkammern der Stenolaemata (Bryozoa). Konstructions-
morphologie und phylogenetische Bedeutung. Cour Forsch Senck.
1991;136:1–263.
32. Jablonski D, Lidgard S, Tayor PD. Comparative ecology of bryozoan
radiations: origin of novelties in cyclostomes and cheilostomes. Palaios.
1997;12:505–23.
33. Ernst A, Schäfer P. Palaeozoic vs post-Palaeozoic Stenolaemata: phy-
logenetic relationship or morphological convergence? Cour Forsch
Senck. 2006;257:49–64.
34. Taylor PD, Michalik J. Cyclostome bryozoans from the Late Triassic
(Rhaetian) of the West Carpathians, Czechoslovakia. N Jahrb Geol P-A.
1991;182:285–302.
35. Boardman RS, McKinney FK, Taylor PD. Morphology, anatomy, and
systematics of the Cinctiporidae, new family (Bryozoa: Stenolaemata).
Sm C Paleob. 1992;70:1–81.
36. Reed CG. Bryozoa. In: Giese AC, Pearse JS, Pearse VB, editors. Reproduc-
tion of marine invertebrates: echinoderms and lophophorates. Pacific
Grove: The Boxwood Press; 1991. p. 85–245.
37. Franzén A. On spermiogenesis, morphology of the spermatozoon,
and biology of fertilization among invertebrates. Zool Bidr Uppsala.
1956;31:355–482.
38. Franzén A. Gametogenesis of bryozoans. In: Wollacott RM, Zimmer RL,
editors. Biology of bryozoans. New York: Academic Press; 1977. p. 1–22.
39. Franzén A. Ultrastructure of spermatids and spermatozoa in the cyclos-
tomatous bryozoan Tubulipora (Bryozoa, Cyclostomata). Zoomorphol-
ogy. 1984;104:140–6.
40. Harmer SF. On the occurrence of embryonic fission in cyclostomatous
Polyzoa. Q J Microsc Sci. 1893;34:199–241.
41. Harmer SF. On the development of Lichenopora verrucaria, Fabr. Q J
Microsc Sci. 1896;39:71–144.
42. Harmer SF. On the development of Tubulipora, and on some British and
northern species of this genus. Q J Microsc Sci. 1898;41:73–157.
43. Robertson A. Embr yology and embryonic fission in the genus Crisia.
Univ Calif Publ Zool. 1903;1:1125–56.
44. Borg F. Studies on recent cyclostomatous Bryozoa. Zool Bidr Uppsala.
1926;10:181–507.
45. Harmer SF. On the origin of the embryos in the ovicells of cyclostoma-
tous Polyzoa. Proc Cam Phil Soc. 1890;7:48.
46. Harmer SF. On the British species of Crisia. Q J Microsc Sci.
1891;32:127–81.
47. Harmer SF. On the development of Lichenopora verrucaria, Fabr. Proc R
Soc Lond. 1895;59:73–4.
48. Harmer SF. Notes on cyclostomatous Polyzoa. Proc Cam Phil Soc.
1896;9:208–14.
49. Waters AW. Bryozoa. Res Voy Belgica Zool. 1904;4:1–114.
50. Waters AW. Bryozoa from Franz-Josef Land, collected by the Jaekson-
Harmaworth Expedition, 1896–1897. Part II. Cyclostomata, Ctenosto-
mata, and Endoprocta. J Linn Soc Zool. 1904;29:161–84.
51. Waters AW. The marine fauna of British East Africa and Zanzibar from
collections made by Cyril Crossland. Bryozoa-Cyclostomata, Ctenosto-
mata and Endoprocta. Proc Zool Soc London. 1914;2:831–58.
52. Borg F. On the structure of cyclostomatous Bryozoa. Ark Zool.
1923;15:1–17.
53. Borg F. A revision of the recent Heteroporidae (Bryozoa). Zool Bidr
Uppsala. 1933;14:253–394.
54. Schwaha T, Handschuh S, Ostrovsky AN, Wanninger A. Morphology of
the bryozoan Cinctipora elegans (Cyclostomata, Cinctiporidae) with the
first data on its sexual reproduction and the cyclostome neuro-muscu-
lar system. BMC Evol Biol. 2018;18:92.
55. Couch RQ. A Cornish fauna, being a compendium of the natural history
of the county. Part 3. The zoophytes and calcareous corallines. Trubo:
Royal Institution of Cornwall; 1844.
56. Smitt FA. Kritisk förteckning öfver Skandinaviens Hafs-Bryozoer. I. Ofvers
Kongl Vetenskap Akad Forh Stockholm. 1865;22:115–42.
57. Smitt FA. Om hafs-bryozoernas utveckling och fettkroppar. Ofvers
Kongl Vetenskap Akad Forh Stockholm. 1865;22:5–50.
58. Barrois J. Recherches sur l’embryologie des Bryozoaires. Lille: Six-
Horemans; 1877.
59. Metschnikoff E. Vergleichend-embryologische Studien. Z Wiss Zool.
1882;37:277–313.
60. Ostroumoff AA. Contribution a l’etude zoologique et morphologique
des bryozoaires du Golfe de Sebastopol. Arch Slav Biol. 1886;1:557–69.
61. Ostroumoff AA. Zur Entwicklungsgeschichte der cyclostomen Seebryo-
zoen. Mitt Zool Stat Neapel. 1887;7:177–90.
62. Calvet L. Contribution `a l’histoire naturelle des Bryozaires Ectoproctes
marins. Trav Inst Zool Univ Montpellier. 1900;8:1–488.
63. Marcus E. Zur vergleichenden Embryologie der Bryozoen. Mitt Zool
Mus Berlin. 1924;11:156–66.
64. Nielsen C. On metamorphosis and ancestrula formation in cyclostoma-
tous bryozoans. Ophelia. 1970;7:217–56.
65. Lutov NP, Minitchev YS. Some peculiarities of development of the
cyclostomatous bryozoans. Zool Zh. 1979;58:789–96 ([in Russian]).
66. Dolinina A. Study of embryonic development of the cyclostome bryo-
zoan Crisiella producta. Unpublished Diploma Thesis, Saint Petersburg
State University; 1992 ([in Russian]).
67. D’Hondt J-L. Structure larvaire et histogenese post-larvaire chez Crisia
denticulata (Lamarck) (Bryozoa, Cyclostomata, Articulata). Zool Scr.
1977;6:55–67.
68. Nielsen C, Bekkouche NT, Worsaae K. Neuromuscular structure of the
larva to early ancestrula stages of the cyclostome bryozoan Crisia
eburnea. Acta Zool. 2019;100:268–81.
69. Jenkins HL, Bishop JDD, Hughes RN. Prudent female allocation by
modular hermaphrodites: female investment is promoted by the
opportunity to outcross in cyclostome bryozoans. Biol J Linn Soc.
2015;116:593–602.
70. Jenkins HL, Waeschenbach A, Okamura B, Hughes RN, Bishop JDD.
Phylogenetically widespread polyembryony in cyclostome bryozoans
and the protracted asynchronous release of clonal brood-mates. PLoS
ONE. 2017;12:e0170010.
71. Hyman LH. The invertebrates. Smaller coelomate groups. New York:
McGraw-Hill; 1959.
72. Santagata S. Ectoprocta. In: Wanninger A, editor. Evolutionary develop-
mental biology of invertebrates. Vol. 2: Lophotrochozoa (Spiralia). Wien:
Springer; 2015. p. 247–62.
73. Nielsen C, Pedersen KJ. Cystid structure and protrusion of the polypide
in Crisia (Bryozoa, Cyclostomata). Acta Zool. 1979;60:65–88.
74. Mukai H, Terakado KY, Reed CG. Bryozoa. In: Harrison FW, editor. Micro-
scopic anatomy of invertebrates, vol. 13. New York: Wiley-Liss; 1997. p.
45–206.
75. Baluška F, Volkmann D, Barlow PW. Eukaryotic cells and their cell bodies:
cell theory revised. Ann Bot. 2004;94:9–32.
76. Hincks T. A history of the British marine Polyzoa. London: Van Voorst;
1880.
77. Robertson A. The cyclostomatous Bryozoa of the west coast of North
America. Univ Calif Publ Zool. 1910;6:225–84.
78. Ostrovsky AN. Comparative studies of ovicell anatomy and reproduc-
tive patterns in Cribrilina annulata and Celleporella hyalina (Bryozoa:
Cheilostomata). Acta Zool. 1998;79:287–318.
79. Ostrovsky AN, Porter JS. Pattern of occurrence of supraneural coe-
lomopores and intertentacular organs in Gymnolaemata (Bryozoa) and
its evolutionary implications. Zoomorphology. 2011;130:1–15.
80. Weedon MJ, Taylor PD. Skeletal ultrastructure in some articulate cyclos-
tome bryozoans. Acta Zool. 1998;79:133–48.
81. Borg F. On the body-wall in Bryozoa. Q J Microsc Sci. 1926;70:583–98.
82. Boardman RS, Cheetham AH. Degrees of colony dominance in stenol-
aemate and gymnolaemate Bryozoa. In: Boardman RS, Cheetham AH,
Oliver WA, editors. Animal colonies. Development and function through
time. Stroudsburg: Dowden Hutchinson and Ross; 1973. p. 121–220.
83. Carle KJ, Ruppert EE. Comparative ultrastructure of the bryozoan
funiculus: a blood vessel homologue. Z Zool Syst Evol. 1983;21:181–93.
84. Taylor PD, Jones CG. Skeletal ultrastructure in the cyclostome bryozoan
Hornera. Acta Zool. 1993;74:135–43.
85. Taylor PD, Weedon MJ, Jones CG. Skeletal ultrastructure in some
cyclostome bryozoans of the family Lichenoporidae. Acta Zool.
1995;76:205–16.
86. Weedon MJ, Taylor PD. Calcitic nacreous ultrastructures in bryozoans:
implications for comparative biomineralization of lophophorates and
molluscs. Biol Bull. 1995;188:281–92.

Page 33 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
87. Weedon MJ, Taylor PD. Skeletal ultrastructures in some cerioporine
cyclostome bryozoans. Acta Zool. 1996;77:249–65.
88. Taylor PD, Weedon MJ. Skeletal ultrastructure and phylogeny of cyclos-
tome bryozoans. Zool J Linn Soc. 2000;128:337–99.
89. Weedon MJ, Taylor PD. Skeletal ultrastructure in some tubuliporine
cyclostome bryozoans. Acta Zool. 1997;78:107–22.
90. Nekliudova UA, Shunkina KV, Grishankov AV, Varfolomeeva MA, Grano-
vitch AI, Ostrovsky AN. Colonies as dynamic systems: reconstructing the
life history of Cribrilina annulata (Bryozoa) on two algal substrates. J Mar
Biol Assoc UK. 2019;99:1363–77.
91. Kvach AY, Varfolomeeva MA, Kotenko ON, Sukhotin AA, Kutiumov VA,
Grischenko AV, Granovitch AI, Ostrovsky AN. Life history of the ctenos-
tome bryozoan Flustrellidra hispida in the White Sea. Invertebr Zool.
2019;16:343–60.
92. Shevchenko ET, Varfolomeeva MA, Nekliudova UA, Kotenko ON, Usov
NV, Granovitch AI, Ostrovsky AN. Electra vs Callopora: life histories of two
bryozoans with contrasting reproductive strategies in the White Sea.
Invertebr Reprod Dev. 2020;64:137–57.
93. Faulkner GH. The relation between somatic and germ cells in the asexu-
ally produced polyps of the polyzoon Alcyonidium gelatinosum. Ann
Mag Nat Hist. 1933;11:255–69.
94. Ostrovsky AN. External versus internal and self- versus cross-. Fertiliza-
tion in Bryozoa: transformation of the view and evolutionary considera-
tions. In: Wyse JPN, Spencer JME, editors. Annals of Bryozoology 2:
Aspects of the history of research on bryozoans. Dublin: International
Bryozoology Association; 2008. p. 103–15.
95. Ostrovsky AN, Taylor PD. Systematics of some Antarctic Idmidronea and
Exidmonea (Bryozoa: Cyclostomata). J Nat Hist. 1996;30:1549–75.
96. Pemberton AJ, Hughes RN, Manrìquez PH, Bishop JDD. Efficient utiliza-
tion of very dilute aquatic sperm: sperm competition may be more
likely than sperm limitation when eggs are retained. Proc R Soc Lond B.
2003;270:223–6.
97. Hughes RN, D’Amato ME, Bishop JD, Carvalho GR, Craig SF, Hansson LJ,
Harley MA, Pemberton AJ. Paradoxal polyembryony? Embryonic clon-
ing in the ancient order of marine bryozoans. Biol Let. 2005;1:178–80.
98. Ryland JS. ‘Polyembryony paradox’: the case of cyclostomate Bryozoa.
Trends Ecol Evol. 1996;11:26.
99. Vines SH. A students’ textbook of botany, vol. 1. London: Swan Sonnen-
schein; 1894.
100. Daubenmire RF. The use of the terms coenocyte and syncytium in biol-
ogy. Science. 1936;84:533.
101. Lee RE. Phycology. Cambridge: Cambridge University Press; 2018.
102. Webster J, Weber R. Introduction to fungi. Cambridge: Cambridge
University Press; 2007.
103. Sullivan W, Theurkauf WE. The cytoskeleton and morphogenesis of the
early Drosophila embryo. Curr Opin Cell Biol. 1995;7:18–22.
104. Fritzenwanker JH, Genikhovich G, Kraus Y, Technau U. Early develop-
ment and axis specification in the sea anemone Nematostella vectensis.
Dev Biol. 2007;310:264–79.
105. Carvalho L, Heisenberg CP. The yolk syncytial layer in early zebrafish
development. Trends Cell Biol. 2010;20:586–92.
106. Scholtz G, Wolff C. Arthropod embryology: cleavage and germ band
development. In: Minelli A, Boxshall G, Fusco G, editors. Arthropod biol-
ogy and evolution. Berlin: Springer; 2013. p. 63–89.
107. Slyusarev GS. Phylum Orthonectida: morphology, biology, and relation-
ships to other multicellular animals. Zhurn Obshc Biol. 2008;69:403–27
([in Russian]).
108. Kent ML, Andree KB, Bartholomew JL, El-Matbouli MA, Desser SS, Devlin
RH, Feist SW, Hedrick RP, Hoffmann RW, Khattra J, Hallett SL, Lester RJG,
Longshaw M, Palenzeula O, Siddall M, Xiao C. Recent advances in our
knowledge of the Myxozoa. J Eukaryot Microbiol. 2001;48:395–413.
109. Boardman RS, Cheetham AH, Cook PL. Introduction to the Bryozoa.
In: Robinson RA, editor. Treatise on invertebrate paleontology Bryozoa
(Part G, revised). Boulder: Geological Society of America, Lawrence:
University of Kansas Press; 1983. p. 3–48.
110. Ostrovsky AN. Variability of ooeciostome shape and position in
Antarctic idmidroniform bryozoans (Bryozoa: Tubuliporida). Zool Anz.
1998;237:97–106.
111. Lutov NP. Some peculiarities of development of the cyclostomatous
bryozoans. Unpublished Diploma Thesis, Saint Petersburg State Univer-
sity; 1977 ([in Russian]).
112. Martín-Durán JM, Egger B. Developmental diversity in free-living flat-
worms. EvoDevo. 2012;3:1–23.
113. Sutton MF. The sexual development of Salpa fusiformis (Cuvier). Part I. J
Embryol Exp Morphol. 1960;8:268–90.
114. Allen WR, Wilsher S, Turnbull C, Stewart F, Ousey J, Rossdale PD, Fowden
AF. Influence of maternal size on placental, fetal and neonatal growth in
the horse. I. Development in utero. Reproduction. 2002;123:445–53.
115. Wilsher S, Allen WR. The effects of maternal age and parity on placental
and fetal development in the mare. Equine Vet J. 2003;35:476–83.
116. Woollacott RM, Zimmer RL. A simplified placenta-like system for the
transport of extraembryonic nutrients during embryogenesis of Bugula
neritina (Bryozoa). J Morphol. 1975;147:355–78.
117. Bishop JDD, Pemberton AJ. The third way: spermcast mating in sessile
marine invertebrates. Integr Comp Biol. 2006;46:398–406.
118. Godeaux JEA. Urochordata—Thaliacea. In: Adiyodi KG, Adiyodi RG, edi-
tors. Reproductive biology of invertebrates Part B: Fertilization, develop-
ment, and parental care. New York: Wiley; 1990. p. 453–69.
119. Bibiermair J, Ostrovsky AN, Wanninger A, Schwaha TF. submitted. The
embryonic development and matrotrophy in the phylactolaemate
bryozoan Plumatella casmiana. Org. Divers. Evol.
120. Ström R. Brooding patterns of bryozoans. In: Woollacott RM, Zimmer RL,
editors. Biology of bryozoans. New York: Academic Press; 1977. p. 23–55.
121. Ostrovsky AN, Nielsen C, Vávra N, Yagunova EB. Diversity of the brood-
ing structures in calloporid bryozoans (Gymnolaemata: Cheilostomata):
comparative anatomy and evolutionary trends. Zoomorphology.
2009;128(1):13–35.
122. Ostrovsky AN, O’Dea A, Rodrígues F. Comparative anatomy of internal
incubational sacs in cupuladriid bryozoans and the evolution of brood-
ing in free-living cheilostomes. J Morphol. 2009;270:1413–30.
123. Marcus E. Sobre o desenvolvimento do bryozoario Synnotum aegyptia-
cum. Arq Cirurg Clin Exp. 1941;5:227–34.
124. Dyrynda PEJ. A preliminary study of patterns of polypide generation-
degeneration in marine cheilostome Bryozoa. In: Larwood GP, Nielsen
C, editors. Recent and fossil Bryozoa. Fredensborg: Olsen and Olsen;
1981. p. 73–81.
125. Pace RM. On the early stages in the development of Flustrella hispida
(Fabricius), and on the existence of a ‘“yolk nucleus”’ in the egg of this
form. Q J Microsc Sci. 1906;50:435–78.
126. Gould SJ, Vrba ES. Exaptation—a missing term in the science of form.
Paleobiology. 1982;8:4–15.
127. Dohrn A. Ursprung der Wirbelthiere und das Princip des Function-
swechsels: genealogische Skizzen. Leipzig: W. Engelmann; 1875.
128. Kleinenberg N. Über die Entwicklung durch Substitution von Organen.
Z Wiss Zool. 1886;44:212–25.
129. Plate L. Deszendenztheorie. Handw Naturwiss. 1912;2:897–951.
130. Plate L. Allgemeine Zoologie und Abstammungslehre. Vol. 2. Die Sin-
nesorgane der Tiere. Jena: G. Fisher; 1924.
131. Fedotov DM. On some cases of morphological and physiological
substitutions. In: Severtzoff AN, Matveiev BS, editors. Proseedings of the
2nd congress of zoologists, anatomists and histologists of USSR, 1925.
Moscow: Glavnauka; 1927. p. 94–5 ([in Russian]).
132. Sewertzoff AN. On the principles of phylogenies. Rev Zool Rus (Rus Zool
Zh). 1928;3:139–48 ([in Russian with German summary]).
133. Sewertzoff AN. Directions of evolution. Acta Zool. 1929;10:59–141.
134. Sewertzoff AN. Morphological regularities of evolution. Moscow, Lenin-
grad: Publishing House of the Academy of Sciences of the USSR; 1939
([in Russian]).
135. Schmalhausen II. Foundations of comparative anatomy. Moscow:
Uchpedgiz; 1938. ([in Russian]).
136. Sewertzoff AS. Directionality of evolution. Moscow: Moscow State
University; 1990 ([in Russian]).
137. Byrne BM. Ophiuroidea. In: Harrison FW, Chia FS, editors. Microscopic
anatomy of invertebrates: Echinodermata, vol. 14. New York: Wiley-Liss;
1994. p. 247–343.
138. Kimura M, Ohta T. On some principles governing molecular evolution.
Proc Natl Acad Sci. 1974;71:2848–52.
139. Ludwig MZ. Functional evolution of non-coding DNA. Cur Opin Gen
Dev. 2002;12:634–9.
140. Fay JC, Wu CI. Sequence divergence, functional constraint, and selec-
tion in protein evolution. Annu Rev Genom Hum G. 2003;4:213–35.

Page 34 of 34
Nekliudovaetal. BMC Ecol Evo (2021) 21:54
•
fast, convenient online submission
•
thorough peer review by experienced researchers in your field
•
rapid publication on acceptance
•
support for research data, including large and complex data types
•
gold Open Access which fosters wider collaboration and increased citations
maximum visibility for your research: over 100M website views per year
•
At BMC, research is always in progress.
Learn more biomedcentral.com/submissions
Ready to submit your research
Ready to submit your research
? Choose BMC and benefit from:
? Choose BMC and benefit from:
141. Ng PC, Henikoff S. SIFT: predicting the effects of aminoacid substitu-
tions on protein function. Annu Rev Genom Hum Genet. 2006;7:61–80.
142. Blackburn DG. Chorioallantoic placentation in squamate rep-
tiles: structure, function, development, and evolution. J Exp Zool.
1993;266:414–30.
143. Wourms JP. Viviparity: the maternal-fetal relationships in fishes. Am
Zool. 1981;21:473–515.
144. Jones MK. Structure and diversity of cestode epithelia. Int J Parasitol.
1998;28:913–23.
145. Korneva ZV. Characterization of Cestoda tissue organization. Biol Bull.
2013;40:146–57.
146. Korneva ZV. Placental type interactions and evolutionary trends
of development of uterus in cestodes. J Evol Biochem Physiol.
2005;41:552–60.
147. Korneva ZV, Kornienko SA. Morphology and ultrastructure of the uterus
of Lineolepis scutigera (Dujardin, 1845) Karpenko, 1985 (Cestoda, Cyclo-
phyllidea, Hymenolepididae) in formation of uterine capsules. Inland
Water Biol. 2013;6:259–67.
148. Korneva JV, Kuklin VV, Kornienko SA. Ultrastructure of the reticulate
uterus and specific features of matrotrophy in three species of higher
cestodes (Cestoda, Cyclophyllidea). Inland Water Biol. 2016;9:234–41.
149. Cable J, Harris PD. Gyrodactylid developmental biology: historical
review, current status and future trends. Int J Parasitol. 2002;32:255–80.
150. Bakke TA, Cable J, Harris PD. The biology of gyrodactylid monogeneans:
the “Russian-doll killers”. Adv Parasit. 2007;64:161–376.
151. Cable J, Harris PD, Tinsley RC. Ultrastructural adaptations for viviparity in
the female reproductive system of gyrodactylid monogeneans. Tissue
Cell. 1996;28:515–26.
152. Jones MK, Ernst I, Whittington ID. The uterine epithelium of Gyrodac-
tylus kobayashii (Monogenea: Gyrodactylidae): ultrastructure of basal
matrices, cytoplasmic membranes and the birth plug, and comparison
with other reproductive epithelia. Int J Parasitol. 1998;28:1805–15.
153. Walker M, Campiglia S. Some aspects of segment formation and
post-placental development in Peripatus acacioi Marcus and Marcus
(Onychophora). J Morphol. 1988;195:123–40.
154. Walker M, Campiglia S. Some observations on the placenta and embry-
onic cuticle during development in Peripatus acacioi Marcus & Marcus
(Onychophora, Peripatidae). In: Minelli A, editor. Proceedings of the
7th international congress of myriapodology. Leiden: EJ Brill; 1990. pp.
449–59.
155. Campiglia SS, Walker MH. Developing embryo and cyclic changes in
the uterus of Peripatus (Macroperipatus) acacioi (Onychophora, Peripati-
dae). J Morphol. 1995;224:179–98.
156. Brooks WK. On the nutrition of the Salpa embryo. Ann Mag Nat Hist.
1893;6:369–74.
157. Bone Q, Pulsford AL, Amoroso EC. The placenta of the salp (Tunicata:
Thaliacea). Placenta. 1985;6:53–63.
158. Berrill NJ. Budding and development in Salpa. J Morphol. 1950;87:553–606.
159. Brien P. Contribution à l’étude de l’embryogenese et de la blastogenese
des Salpes. Rec Inst Zool Torley-Rousseau. 1928;2:5–116.
160. Blackburn DG, Vitt LJ. Specializations of the chorioallantoic placenta in
the Brazilian scincid lizard, Mabuya heathi: a new placental morphotype
for reptiles. J Morphol. 2002;254:121–31.
161. Ramirez-Pinilla MP, De Perez G, Carreno-Escobar JF. Allantoplacen-
tal ultrastructure of an Andean population of Mabuya. J Morphol.
2006;267:1227–47.
162. Korneva JV, Kornienko SA, Guljaev VD. Fine structure of the gravid uteri
of two tapeworm species Skrjabinacanthus diplocoronatus Spassky et
Morosov, 1959 and Urocystis prolifer Villot, 1880 (Cestoda: Cyclophyl-
lidea) parasitic in shrews that display different fecundity of the strobilae.
Parasitol Res. 2012;111:1523–30.
163. Ereskovsky AV. The comparative embryology of sponges. London:
Springer; 2010.
164. Ivanova-Kazas OM. Asexual propagation of animals. Leningrad: Lenin-
grad State University; 1977. ([In Russian]).
165. Craig SF, Slobodkin LB, Wray G. The ‘paradox’ of polyembryony: a
review of the cases and a hypothesis for its evolution. Evol Ecol.
1997;11:127–43.
166. Grbic M. Polyembryony in parasitic wasps: evolution of a novel mode of
development. Int J Dev Biol. 2003;47:633–42.
167. Pemberton AJ, Hansson LJ, Craig SF, Hughes RN, Bishop JDD. Microscale
genetic differentiation in a sessile invertebrate with cloned larvae:
investigating the role of polyembryony. Mar Biol. 2007;153:71–82.
168. Borg F. The stenolaematous Bryozoa. Further Zool Results Swed Antarc
Exped. 1901–1903. 1944;3:1–276.
169. Harmelin JG. Le sous-ordre des Tubuliporina (Bryozoaires Cyclostomes)
en Mediterranée. Ecologie et Systematique. Mem Inst Oceanogr,
Monaco. 1976;10:1–326.
170. Wood ACL, Probert PK, Rowden AA, Smith AM. Complex habitat
generated by marine bryozoans: a review of its distribution, structure,
diversity, threats and conservation. Aq Conserv Mar Freshwat Ecos.
2012;22:547–63.
171. Iwabuchi K. Polyembryonic insects. Singapore: Springer; 2019.
172. Linnaeus C. Systema naturae. Holmiae, Salvii; 1758
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in pub-
lished maps and institutional affiliations.