ArticlePDF Available

Taphonomy of insects in carbonates and amber


Abstract and Figures

The major taphonomic processes that control insect preservation in carbonate rocks (limestones, travertines and nodules) are biological: insect size and wingspan, degree of decomposition, presence of microbial mats, predation and scavenging; environmental: water surface tension, water temperature, density and salinity, current activity; and diagenetic: authigenic mineralisation, flattening, deformation, carbonisation. The major taphonomic processes that control the preservation of insects in fossil resins (amber and copal) are different, but can be considered under the same headings – biological: presence of resin producers, size and behaviour of insects; environmental: latitude, climate, seasonality, resin viscosity, effects of storms and fires, soil composition; and diagenetic: resin composition, insect dehydration, pressure, carbonisation, thermal maturation, reworking, oxidation. These taphonomic processes are geographically and temporally restricted, and generate biases in the fossil record. Nevertheless, where insects occur they may be abundant and very diverse. Taphonomic processes may impact on phylogenetic and palaeobiogeographic studies, in determining the timing of the origin and extinction of insect groups, and in identifying radiations and major extinctions. Taphonomic studies are an essential prerequisite to the reconstruction of fossil insect assemblages, to interpreting the sedimentary and environmental conditions where insects lived and died, and to the investigation of interactions between insects and other organisms.
Content may be subject to copyright.
Taphonomy of insects in carbonates and amber
Xavier Mart|
'sa, Derek E.G. Briggs b;, Enrique Penalver c
aDept. Estratigra¢a, Paleontologia i Geocie
'ncies Marines, Fac. Geologia, Universitat de Barcelona, 08028 Barcelona, Spain
bDepartment of Geology and Geophysics, Yale University, P.O.Box 208109, New Haven, CT 06520-8109, USA
cInstitut Cavanilles de Biodiversitat i Biologia evolutiva, Universitat de Vale
'ncia, Apart. 2085 Vale
'ncia, Spain
Received 15 July 2002; received in revised form 11 August 2003; accepted 16 September 2003
The major taphonomic processes that control insect preservation in carbonate rocks (limestones, travertines and
nodules) are biological: insect size and wingspan, degree of decomposition, presence of microbial mats, predation and
scavenging; environmental: water surface tension, water temperature, density and salinity, current activity; and
diagenetic: authigenic mineralisation, flattening, deformation, carbonisation. The major taphonomic processes that
control the preservation of insects in fossil resins (amber and copal) are different, but can be considered under the
same headings ^ biological: presence of resin producers, size and behaviour of insects; environmental : latitude,
climate, seasonality, resin viscosity, effects of storms and fires, soil composition; and diagenetic: resin composition,
insect dehydration, pressure, carbonisation, thermal maturation, reworking, oxidation. These taphonomic processes
are geographically and temporally restricted, and generate biases in the fossil record. Nevertheless, where insects occur
they may be abundant and very diverse. Taphonomic processes may impact on phylogenetic and palaeobiogeographic
studies, in determining the timing of the origin and extinction of insect groups, and in identifying radiations and
major extinctions. Taphonomic studies are an essential prerequisite to the reconstruction of fossil insect assemblages,
to interpreting the sedimentary and environmental conditions where insects lived and died, and to the investigation of
interactions between insects and other organisms.
 2003 Elsevier B.V. All rights reserved.
Keywords: fossil insects ; preservation; limestone ; resin
1. Introduction
Insects are by far the most diverse and success-
ful group of macroscopic organisms and they play
an important role in all the terrestrial ecosystems
that they inhabit. The fossil record of insects and
their diversity through time have been reviewed
(Carpenter, 1992; Labandeira and Sepkoski,
1993; Jarzembowski and Ross, 1996; Ross et
al., 2000; Jarzembowski, 2001b) as well as the
palaeobiology of insect feeding (Labandeira et
al., 1994; Labandeira, 1997). The earliest record
of the Insecta is from the Lower Devonian of
¤bec) (Labandeira et al., 1988), but
the group is not evident in the fossil record in
signi¢cant numbers until the Upper Carbonifer-
ous (Brauckmann et al., 1995). The number of
0031-0182 / 03 / $ ^ see front matter 2003 Elsevier B.V. All rights reserved.
doi: 10.1016/S0031-0182(03)00643-6
* Corresponding author. Tel./Fax: +1-203-432-8590.
E-mail addresses:
(X. Mart|
's), (D.E.G. Briggs), (E. Penalver).
PALAEO 3225 5-1-04
Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^64
insect orders present in the Permian was similar to
that of now; the record of families identi¢es ma-
jor periods of origination in the Permo^Carbon-
iferous, Early Jurassic, Early Cretaceous and Pa-
laeocene (Jarzembowski and Ross, 1996). The
degree to which the recorded diversity of insects
is a re£ection of taphonomic processes is very
di⁄cult to determine. A knowledge of the condi-
tions that led to the preservation of insect biotas,
and of the biases introduced by taphonomic pro-
cesses, is also essential for interpreting the role of
insects in terrestrial ecosystems, such as organic
matter recycling and the pollination and distribu-
tion of plant taxa. Here we review the major ta-
phonomic processes that a¡ect insects.
Insects lack biomineralised tissues and are usu-
ally considered by palaeontologists as soft-bodied
organisms. Exceptional conditions are normally
required to ensure their preservation, but where
taphonomic processes are favourable, insects may
be extremely abundant. However, there is clearly
a signi¢cant range in susceptibility to decay, for
example between £ies and beetles, re£ecting con-
trasts in the degree of sclerotisation of the cuticle.
Nonetheless, laboratory experiments have shown
that even £y carcasses may survive in quiet sedi-
mentary settings for more than a year without
disarticulating (Mart|
's and Martinell,
The most exceptionally preserved insects in
sedimentary rocks occur in ¢ne-grained laminated
carbonates in lacustrine and shallow marine set-
tings, where conditions may be suitable for the
preservation of truly labile soft-tissues (e.g. Soln-
hofen, Germany: Frickhinger, 1994). They pro-
vide a basis for contrasting preservation with
that in amber. The di¡erent taphonomic processes
that control the preservation of insects in carbon-
ate rocks and in amber result in samples of di¡er-
ent insect communities. Amber normally preserves
insects regardless of their susceptibility to decay,
but selective preservation is often a feature of in-
sect assemblages in carbonates. The occurrence of
insects is normally considered to identify a fossil
deposit as a Konservat-Lagersta
«tte or conserva-
tion deposit (sensu Seilacher, 1970) where the em-
phasis is on the quality of preservation rather
than the abundance of fossils. Konservat-Lager-
«tten which preserve non-biomineralised tissues
are a critical source of palaeobiological data that
are not available from the ‘shelly’ fossil record
(Briggs, 1995a). The preservation of delicate
structures allows for more detailed comparison
with recent organisms. The controls on the pres-
ervation of non-biomineralised tissues, so-called
soft-bodied fossils, are more complex than those
on shelly taxa, which are dominantly sedimento-
logical (Kidwell, 1991) ^ they include organic
matter input, microbial activity, and environmen-
tal geochemistry (Allison and Briggs, 1991a,b ;
Briggs, 2003a).
Insect body fossils occur in a variety of envi-
ronmental settings including peat deposits (Ken-
ward, 1976; Hayashi, 1994; Lavoie et al., 1997),
deserts (e.g. Chihuahuan Desert, USA, Quater-
nary: Elias, 1990; desert sands Mauritania, Qua-
ternary, Azar pers. commun.), lakes and rivers
(e.g. Montsec and Las Hoyas, Spain, Lower
Cretaceous: Mart|
's, 1995; Mele
1995; Se
¤zanne and Auriol travertines, France, Eo-
cene: Nel and Blot, 1990; Papazian and Nel,
1989), deltas (e.g. Vosges, France, Middle Trias-
sic: Gall, 1996; Marchal-Papier, 1998), lagoons
(e.g. Solnhofen, Germany, Upper Jurassic :
Malz, 1976; Frickhinger, 1994), open marine en-
vironments (e.g. northern Switzerland, Lower Ju-
rassic: Etter and Kuhn, 2000), and deeper marine
turbidites (e.g. Borreda
', Spain, Eocene: Gaudant
and Busquets, 1996). The spatiotemporal distribu-
tion of these palaeoenvironments was a major
control on the insect fossil record.
Where insects are preserved in calcium carbon-
ate, precipitation may occur as calcite mud (e.g.
Green River, USA, Eocene: Ferber and Wells,
1995), aragonite mud (e.g. Rubielos de Mora,
Spain, Miocene: Pen‹alver, 1998), and dolomite
mud (e.g. Karatau, Kazakhstan, Upper Jurassic :
Seilacher et al., 1985). Fossil insects also are pre-
served in a variety of other sedimentary contexts :
clays and marls (Nel, 1986), siltstones (Mart|
's and Nel, 1991), sandstones (Nel et al.,
1993), lacustrine diatomites (Hong, 1985 ; Riou,
1995), cherts (Whalley and Jarzembowski, 1981),
evaporites (Priesner and Quievreux, 1935 ; Schlu
ter and Kohring, 2001), phosphates (Handschin,
1944), coal measures (Bartram et al., 1987 ; Shear
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^6420
and Kukalova
¤-Peck, 1990; Jarzembowski, 2001a),
and asphalt (Miller, 1983, 1997; Iturralde-Vinent
et al., 2000). Here we focus on the two most im-
portant sources of fossil insects, carbonates and
Repositories are abbreviated as follows :
MNHN, Muse
¤um National d’Histoire Naturelle,
Paris, France; PIN, Palaeontological Institute,
Russian Academy of Sciences, Moscow, Russia ;
MCAM, Museu de la Cie
'ncia, Fundacio
Caixa’, Barcelona, Spain; EPGM, Dept. Estrati-
gra¢a, Paleontologia i Geocie
'ncies Marines, Univ.
Barcelona, Barcelona, Spain; MCNA, Museo
Ciencias Naturales de A
Łlava, Vitoria^Gasteiz,
Spain; MCCM, Museo de Ciencias de Castilla^
La Mancha, Cuenca, Spain; MPV, Museu Pale-
'gic de Vale
'ncia, Vale
'ncia, Spain; JME,
Jura-Museum, Eichsta
«tt Germany; QM, Queens-
land Museum, Brisbane, Australia ; IEI, Institut
d’Estudis Ilerdencs, Lleida, Spain.
2. The importance of insect
«tten reveal the diversity of
insects in the past. Di¡erent taphonomic process-
es bias the preservation of insects in carbonates
and amber in di¡erent ways. A more diverse
range of taxa and sizes is preserved in carbonates
whereas amber is usually dominated by particular
taxonomic groups and size categories (Zherikhin
et al., 1999). In the Lower Cretaceous of Spain,
for example, only a few families of insects occur
in both the Montsec limestones (Pen‹alver et al.,
1999) and A
Łlava amber (Mart|
's et al.,
1999; Alonso et al., 2000). The occurrence of in-
sect compression fossils in sediment that also
yields insect-bearing resin is extremely rare. Nota-
bly, a well preserved cockroach wing was found
associated with amber from the Palaeocene Due
Formation at Urtuy, on the Naiba River of Sa-
khalin Island, Russia (Zherikhin, pers. commun.,
2001). Fossil insects also have been found in both
limestone and amber from the same exposure.
Such amber localities include the Lower Creta-
ceous of Lebanon (Azar, 2000) and the Upper
Cretaceous of New Jersey (Grimaldi et al., 2000).
Evidence of ecologic relationships between in-
sects and other animals is sometimes preserved in
limestones, where it is usually con¢ned to plant^
insect interactions (Labandeira, 1998 ; Waggoner,
1999). Such behaviour is inferred on the basis of
functional morphology, or gut contents (Schaal
and Ziegler, 1992; Krassilov et al., 1997)or
more commonly coprolites (Rothwell and Scott,
1988), or an array of primary evidence for insect
feeding (Ro
«Mler, 2000; Labandeira and Phillips,
2002). Amber, on the other hand, commonly re-
veals interactions, such as reproduction (mating,
egg laying), commensalism and parasitism be-
tween di¡erent insects, and between insects and
other organisms such as nematodes, spiders, pseu-
doscorpions, mites and ticks, and vertebrates
(Poinar, 1984; Poinar et al., 1994; Grimaldi,
1996; Weitschat and Wichard, 1998). Plant dam-
age is very rarely recorded in amber (e.g. Poinar
and Brown, 2002).
«tten reveal other features of
an ecosystem. The presence of certain taxa, such
as termites (Nel and Paicheler, 1993), or assem-
blages of insects, may indicate particular climatic
conditions (McCobb et al., 1998 ; Poinar et al.,
1999; Duringer et al., 2000; Miller and Elias,
2000). Adult and worker termites have been
found in the Lower Cretaceous of Spain and Si-
beria indicating that social behaviour had evolved
in this group by this time, earlier than it did in
ants or bees (Mart|
's and Martinell,
1995). Some taxa imply the presence of others.
Rasnitsyn (1968) described the oldest cephid saw-
£ies (Hymenoptera, Cephidae) from the Lower
Cretaceous of Baissa in Transbaikalia. Cephid
saw£ies are associated exclusively with angio-
sperms and, as predicted, angiosperm leaves and
seeds subsequently were found at the locality
(Rasnitsyn, pers. commun.). The occurrence in
the Lower Cretaceous ambers of Spain of wasps
(Evaniidae), for example, which are parasitic on
the eggs of cockroaches, suggests that cock-
roaches were present even though they have not
been recorded.
«tten provide evidence of
factors controlling the growth of authigenic min-
erals in sediments and amber. In carbonate rocks
fossil insects are usually preserved as organic re-
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^64 21
mains of the cuticle, or as a mould where the
cuticle has been lost during diagenesis or weath-
ering (Mart|
's et al., 1995). Early min-
eralisation may replicate insect morphology in
calcite (McCobb et al., 1998 ; X.M.-D., pers. ob-
servation in Las Hoyas), pyrite transformed to
goethite (Grimaldi and Maisey, 1990), or in cal-
cium phosphate (Duncan et al., 1998). 3-D pres-
ervation may occur in calcium sulphate, such as
gypsum from the Miocene of Alba, Italy (Chiam-
bretti and Damarco, 1993; Schlu
«ter and Kohring,
2001). Authigenic minerals rarely form in resin.
Pyrite may penetrate amber along fracture planes
(Karwowski and Matuszewska, 1999), coating
the insect inclusion (Schlu
«ter and Stu
1982; Krzeminska et al., 1992 ; Grimaldi et al.,
Exceptionally preserved deposits may yield an-
cient biomolecules. DNA has been reported from
insects in amber (Desalle et al., 1992; Cano et al.,
1992, 1993) but its preservation is controversial
due to its susceptibility to both hydrolytic and
oxidative damage (Lindahl, 1993; Smith and Aus-
tin, 1997; Austin et al., 1997), and to the di⁄culty
of eliminating contamination by recent DNA. De-
cay-resistant tissues, such as the cuticles of some
insects, have a higher preservation potential that
is controlled by three major factors (Briggs,
1999): (1) the nature and composition of the cu-
ticle, exempli¢ed by its better preservation in the
thick sclerotised elytra of beetles than in other
insects; (2) the depositional environment, such
as resin vs. sediment, which in£uences decay
rate; and (3) diagenetic history, including poly-
merisation, which is controlled by thermal e¡ects
and reaction time. The chitin and protein in insect
cuticle have only been reported from later Ceno-
zoic deposits (Miller et al., 1993; Stankiewicz et
al., 1998a; Flannery et al., 2001); weevil cuticle in
the Oligocene lacustrine shales of Enspel, Ger-
many, preserves the oldest traces (Stankiewicz et
al., 1997b). In older examples the original biomol-
ecules are altered to kerogen (Briggs and Eglin-
ton, 1994; Stankiewicz et al., 1997a, b ; Briggs,
1999). Stankiewicz et al. (1998b) demonstrated
that chitin is not preserved in Dominican amber,
con¢rming that the preservation of DNA is highly
improbable (Austin et al., 1997).
3. The role of microbial mats in insect preservation
Microbial mats may facilitate the preservation
of insects in a number of ways (Briggs, 2003b).
Mats are complex communities of photosynthetic
prokaryotes (cyanobacteria), diverse unicellular
algae, and chemautotrophic micro-organisms
(Gall, 1990). Anaerobic and aerobic species coex-
ist. Cyanobacteria may be spherical or ¢lament-
like and they form a mat by secreting mucilage.
Details of fossilised microbial mats may be re-
vealed by scanning electron microscopy (Gall et
al., 1985).
Where microbial mats form on the surface of
standing bodies of water, they may trap insect
carcasses and transport them to the sediment^
water interface when the mat sinks (Gall, 1995 ;
Harding and Chant, 2000). Where a carcass is
overgrown by a mat, it is protected from erosion
and from scavengers and burrowing animals
(Gall, 1990, 1995) and is prevented from £oating.
Microbial mats may reduce decay by acting as a
barrier and promoting conditions unfavourable to
certain bacteria (Gall, 1990, 2001). Microbial
mats may also prevent the transfer of ions, lead-
ing to concentrations within the carcass su⁄cient
to promote mineralisation of soft-tissues or of the
mat itself (Briggs and Kear, 1993; Gall et al.,
1994; Sagemann et al., 1999). Where the mat be-
comes mineralised, as in fossils from the Messel
Shale and from Enspel (Wuttke, 1983; Toporski
et al., 2002), it may form a pseudomorph of the
carcass, preserving the gross morphology of the
Microbial mats develop today in extreme envi-
ronments, such as sabkhas, intertidal £ats, and
anoxic marine bottom water, conditions that are
hostile to most organisms. Marine microbial mats
have been reported in the Upper Jurassic Solnho-
fen Limestone, Germany, which yields a signi¢-
cant assemblage of insects (Keupp, 1977; Frick-
hinger, 1994). Most of these insects are preserved
as moulds or by precipitation of calcite or pyro-
lusite; phosphate mineralisation is rare, amount-
ing to 68% of individuals (Wilby et al., 1995).
Lacustrine microbial mats have been reported in
the Miocene basins of Rubielos de Mora and Bi-
corp, Spain, which yield important assemblages of
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^6422
insects (Pen‹alver et al., 2002a). 3-D caddis£y pu-
pae are preserved in Miocene freshwater lime-
stones in Saint-Ge
¤rard-le-Puy, France, associated
with possible mats (Hugueney et al., 1990). How-
ever, more decay prone soft-tissues may not be
mineralised even where microbial mats occur, as
in the Middle Triassic Gre
'Voltzia, Vosges,
France, which yields a diversity of insects (Gall,
1990; Marchal-Papier, 1998). Additionally, in-
sects have not been reported from the Upper Ju-
rassic of Cerin, France, which preserves phospha-
tised soft-tissues in association with microbial
mats (Gall et al., 1985; Wilby et al., 1996).
4. Resin as a preservational medium
Resins are produced by specialised tree cells
and exuded through ¢ssures. Amber is a fossilised
natural resin with properties similar to amor-
phous polymeric glass (Poinar, 1992). Resins are
a complex mixture of terpenoid and/or phenolic
compounds (Anderson and Crelling, 1995). Their
chemical composition is diverse but they are solu-
ble in alcohol and insoluble in water. Terpenoids
may be volatile, where mono- and sesquiterpenes
provide £uidity and act as plasticisers, or they
may be non-volatile, as in the case of diterpenoids
or sometimes triterpenoids (Langenheim, 1995).
Among terpenoids, the most common are oxygen-
ated terpenes: acids, alcohols and esters secreted
from plant parenchyma cells. The polymerisation
of non-volatile terpenoids promotes copal and
amber formation, as volatile terpenoids escape
to the atmosphere. The chemical composition of
amber is only partially known, due to its insolu-
bility. Notably, Lebanese amber has been dis-
solved with chloroforms, but the chemical compo-
sition has not been published (Azar, 1997).
Infrared Spectrometry (IRS) has been employed
in comparative studies of fossil and modern res-
ins, allowing the tree producers to be identi¢ed
(Beck, 1999; Kosmowska-Ceranowicz, 1999),
and the categorisation of amber into di¡erent
types. A classi¢cation of fossil resins was pro-
posed by Anderson and Crelling (1995).
Oils, oleo-resins and resins are produced by
both gymnosperms and angiosperms for defence
against herbivores and pathogens such as insects
and fungi. They are common and diverse in recent
tropical ecosystems (Farrell et al., 1991). Neither
oils nor oleo-resins are known to be a source of
amber (Langenheim, 1995). Copals are resins with
a low oil content, mainly from the araucariacean
Agathis (Philippines, New Zealand) and the legu-
minoseans Hymenaea and Copaifera (Africa, Ca-
ribbean and South America), used to make hard
and elastic varnishes. Some dipterocarpaceans
from Southeast Asia produce dammar resins. Fos-
sil copal of both araucariacean and leguminosean
origin with insect inclusions is known from sev-
eral localities in the Southern Hemisphere (Poi-
nar, 1991a; Schlu
«ter, 1993). Araucariacean copal
occurs in northern New Zealand and Victoria,
Australia, where it is known as kauri, and in the
South Paci¢c region, Indonesia and the Philip-
pines, where it is known as Manila copal. Legu-
minosean copal occurs mainly in Africa, in the
Caribbean and South America. Copal from Hy-
menaea spp. is known from northwestern Mada-
gascar, in the eastern parts of Kenya and Tanza-
nia, the Santander region in Colombia, Minas
Gerais in Brazil, and the eastern part of the Do-
minican Republic. Copals produced by Copaifera
are known from West Africa, and that produced
by dipterocarpaceans from Malaysia and Suma-
tra, where it is known as dammar. Insects and
other arthropods are abundant in copals, partic-
ularly in kauris from New Zealand and copals
from Madagascar, Kenya, Tanzania and Colom-
bia. Due to the young age of these resins, which
range up to a thousand years old, the entombed
insect associations are similar to those of today
«ter, 1993).
Resin is produced by at least three families of
conifers and twelve of angiosperms, but only
some of these generate amber in the fossil record.
The conifers are Pinaceae, Araucariaceae and
Taxodiaceae (Taxodiaceae is now included within
Cupressaceae: Stefanovic et al., 1998); the angio-
sperms are Leguminosae, Burseraceae, Diptero-
carpaceae, Hamamelidaceae (Langenheim, 1995),
and Combretaceae (Nel et al., in press). The Le-
guminoseae include the Southern Hemisphere Hy-
menaea which yields copious quantities of resin
(Langenheim, 1995). Dominican and Mexican
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^64 23
ambers were produced by an extinct relative of
the West Indian locust (H. protera)(Poinar,
1991b). Most Mesozoic ambers and Eocene Baltic
amber are considered to be a product of araucar-
iacean conifers as shown by their similarity to the
resin of Recent Agathis (Langenheim, 1995; Beck,
1999). Anderson and LePage (1995) suggested,
however, that Baltic amber originates from a pi-
naceous conifer similar to Pseudolarix.
5. Distribution of amber through time
It is di⁄cult to produce a comprehensive list of
insect occurrences in carbonates due to the enor-
mous number of localities involved and the range
in quality of preservation and assemblage diver-
sity (Allison and Briggs, 1991b). Amber deposits,
on the other hand, are geographically and tempo-
rally restricted (Poinar, 1992; Grimaldi, 1996).
The earliest known fossilised resins are from
Upper Carboniferous pteridosperms of England
(van Bergen et al., 1995), but amber did not be-
come abundant until the Early Cretaceous with
the rise of the coniferous Araucariaceae, particu-
larly in tropical and subtropical forests. The ear-
liest ambers with inclusions are from Jezzine and
Hammana, in the Lower Cretaceous of Lebanon.
More than 60% of the insects in amber from
Hammana are £ies, including taxa that indicate
Fig. 1. Palaeogeographic maps showing the principal localities (see Appendix 1) that yield amber. (A) Lower Cretaceous. (B)
Upper Cretaceous. (C) Palaeocene. (D) Eocene. (E) Oligocene. (F) Miocene.
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^6424
that the resin formed in a warm climate within a
wet leafy forest (Azar, 2000). While some of
the earliest Cretaceous amber-bearing deposits
(Fig. 1A) formed between the equator and 4‡N
(Israel and Lebanon), they are concentrated in
northern mid-latitudes, between 29‡N (Azerbai-
jan) and 50‡N (Japan), the moist megathermal
zone based on vegetation distributions of Morley
(2000). By the end of the Early Cretaceous, while
occupying a similar equatorial range, amber ex-
tended north from 27‡N (Spain) to near 70‡N
(Khatanga, Russia), coinciding with the moist
megathermal zone of the Northern Hemisphere.
Galle (2000) regarded the climate of the Tethys
Region, where a large number of amber-bearing
localities are to be found, as generally humid dur-
ing the Lower Cretaceous and becoming more
arid in the Upper Cretaceous. The presence of
amber in extensive lower-latitude coals in the
Lower Cretaceous of Lebanon and Spain is con-
sistent with very humid conditions (Barro
¤n et al.,
2001). A small number of localities, such as those
in northern Siberia, represent cool temperate con-
ditions. Equatorial occurrences, corresponding to
hot, relatively dry conditions, occur in Brazil and
the Middle East. Only one amber locality is
known in the Southern Hemisphere, in South
Africa (Gomez et al., 2002).
No Upper Cretaceous amber is known from the
Southern Hemisphere (Fig. 1B). The only equato-
rial occurrence is from the earliest Upper Creta-
ceous of Burma, which predates the widespread
development of megathermal angiosperm-rich for-
ests in equatorial latitudes towards the end of the
Cretaceous (Morley, 2000). Localities are mainly
in the northern, moist, megathermal zone along
the northern margin of the Tethys Ocean and in
North America. Barron and Peterson (1990) con-
sidered that the Tethys Ocean would have been
dominated by two clockwise gyres of ocean sur-
face currents during the Cretaceous, with a dom-
inantly easterly £ow along its the northern side.
Climate modelling (Barron et al., 1989) predicted
a strongly developed monsoon in the Tethys area
during the Cretaceous. Elevated temperatures re-
sulted in high rates of evaporation and precipita-
tion, which produced very high seasonal rainfall
on the northern and southern borders of the Te-
thys Ocean (Barron et al., 1989), allowing the for-
mation of coals (Parrish et al., 1982). In North
America amber localities border the epicontinen-
tal seaway £anked by warm temperate wet vege-
tation (biome 5 of Horrell, 1991), which includes
Cupressaceae. The insect amber assemblage from
the Upper Cretaceous Raritan Formation of New
Jersey (Grimaldi et al., 2000) suggests a warm
temperate or subtropical climate, similar to that
of Siberian and Canadian Upper Cretaceous am-
ber forests. Grimaldi et al. (2000) estimated a pa-
laeolatitude of 32‡N for this amber, but palaeo-
geographic considerations place it at 40‡N.
Greenhouse conditions, with particularly high
temperatures during the Middle Cretaceous (Hu-
ber et al., 2002) are evidenced by the nature of the
terrestrial vegetation which indicates a cool tem-
perate regime even at high latitudes. In this con-
text, ambers are found at s80‡N in Alaska.
Berner (1990) considered that levels of atmospher-
ic CO2were eight to ten times those of the present
day, reaching a maximum at the beginning of the
Late Cretaceous.
The climatic history of the Cenozoic can be di-
vided into two episodes (Pickering, 2000) : the Pa-
laeocene to Eocene interval was characterised by a
continuation of the greenhouse conditions of the
Upper Cretaceous whereas cooling occurred dur-
ing the Oligocene to Recent period with occasion-
al short and warmer episodes. There is a paucity
of Palaeocene amber (Fig. 1C) : only 0.5 major
occurrences per million years, compared to 1.65
in the Upper Cretaceous and nearly 2 in the Eo-
cene. There also are no Palaeocene occurrences in
the Southern Hemisphere nor in the tropical rain
forests of the equatorial Palmae Province. The
few amber occurrences are in the northern rain
forests of the Boreotropical Province, or in tem-
perate latitudes further to the north.
The number and geographical spread of amber
localities is much more extensive during the Eo-
cene (Fig. 1D). Amber is represented in the trop-
ical rain forests of the Southern Megathermal
Province (Argentina) and the African Province
(Nigeria) but the majority of localities (including
Baltic amber) occur in the Boreotropical Prov-
ince. Amber is present in the dry subtropics of
China and the Kamchatka Peninsula to the south
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^64 25
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^6426
of the Boreotropical Province and in more tem-
perate latitudes to the north.
Following the terminal Eocene cooling event
the global extent of rain forests was substantially
reduced. Tropical rain forest remained in the Ca-
ribbean region, as evidenced by Dominican Re-
public and Mexican amber derived from Hyme-
naea, and apart from localities in Tunisia and
Sicily where the climate was presumably dry,
most Oligocene amber is found in a mid-latitude
belt across Eurasia (Fig. 1E). Oligocene amber is
unknown from the Southern Hemisphere, but the
mid-latitude belts and tropical rain forests have
yielded amber during the Miocene in both the
Northern and Southern Hemispheres (Fig. 1F).
Most Mesozoic ambers and Eocene Baltic amber
are a product of araucariacean conifers (but see
Anderson and LePage, 1995). As the climatic con-
straints on the distribution of these taxa are
poorly known, they cannot be used directly to
infer the climatic controls on amber.
6. Insect taphonomy
Taphonomy deals with the incorporation of or-
ganic remains into sediments or other contexts, such
as resin, and the fate of these materials after burial.
It is normally divided into: necrolysis, referring to
death and its causes; biostratinomy, involving the
sedimentary history of the remains prior to buri-
al; and diagenesis, comprising physical and chem-
ical modi¢cations within the sediment or resin.
6.1. Necrolysis
6.1.1. Necrolysis in aquatic settings
Where fossil insects occur in carbonates, bio-
mineralised and non-biomineralised remains of
other organisms are found in association, and
often soft-tissues or delicate structures are pre-
served, as in the Lower Cretaceous of Liaoning,
China (Fig. 2A). Insect assemblages preserved in
carbonate are often dominated by species that
rely on water for ecological reasons such as hab-
itat, for hunting, or for laying eggs. Assemblages
of may£ies, termites, and £ying ants, for example,
may be the result of a mass mortality, and/or
show a bias toward a particular size range. Smith
(2000), for example, demonstrated that the insect
assemblage that accumulated around a recent
ephemeral lake in Arizona is biased towards
smaller robust species, and also to ground-dwell-
ing forms. She compared the diversity of living
insects, mainly beetles, that occur in di¡erent en-
vironments around the lake with the assemblage
preserved in the shallow, subsurface sediments
along the lakeshore. 65% of living beetle families
and 28% of living beetle genera were represented
in the sediments; 100% of the families and 91% of
the genera found dead were present in the live
fauna. The relative abundance of beetle families
in the living assemblage is signi¢cantly di¡erent
from their relative abundance in the sediments,
from which the best represented groups are the
families of hymenopterans (62%) and coleopter-
ans (30%).
Soft-tissues of all organisms are quickly de-
graded by bacteria and fungi, which may com-
pletely destroy a carcass in a few days, depending
on its size, the water temperature and other am-
bient factors. Mart|
's and Martinell
(1993) studied the death of insects in aquatic en-
vironments to determine the taphonomic process-
es that in£uence insect fossilisation. Observations
were made on cockroaches, crickets, earwigs, ter-
Fig. 2. (A) Bellabrunetia catherinae, terrestrial dragon£y with aquatic conchostracans (white arrows). MNHN-LP-R 55232a,
Upper Jurassic/Lower Cretaceous, Liaoning (China); photo by A. Nel; scale bar = 10 mm. (B) Coprolite or regurgitate with drag-
on£y wings. PIN 2904/3 ; Upper Jurassic, Karatau (Kazakhstan) ; scale bar = 10 mm. (C) Homopteran showing wings and body
movements. MCAM0043; Miocene amber from Hispaniola (Dominican Republic); scale bar = 5 mm. (D) Isolated insect mandi-
ble. MCAM0005; Miocene amber, Hispaniola ; scale bar = 0.5 mm. (E) Isolated ant head. MCAM0095; Miocene amber, Hispa-
niola; scale bar = 1 mm. (F) Worker termite with an air bubble extending from its body due to the activity of gut microbes.
MCAM0508; Miocene amber, Hispaniola; scale bar = 2 mm. (G) Female of Diptera, Keroplatidae laying eggs during entrapment,
EPGM-RD-0048; Miocene amber, Hispaniola ; scale bar = 1 mm. (H) Copal stalactites, Mart|
¤n-Closas collection, EPGM; Holo-
cene, Madagascar ; scale bar = 10 mm.
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^64 27
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^6428
restrial bugs, may£ies, true £ies, wasps, ants, bee-
tles, snake£ies, lacewings, and butter£ies. Three
di¡erent stages were considered : (a) on the water
surface, during death and while the body was
£oating, (b) during sinking, while the insect was
alive or immediately following death, and (c) on
the bottom at the substrate^water interface. They
compared their experimental results with fossil in-
sects preserved in lithographic limestones.
The results of this study indicated that di¡erent
types of insects behave in varied ways when they
enter water. There are two patterns.
(1) Non-winged insects show two types of be-
haviour when they fall into water. Those with a
density less than water, such as worker ants, walk
easily on the surface allowing them to escape.
They raise the abdomen with rhythmical move-
ments presumably allowing them to breathe.
Larger wingless forms, such as earwigs, penetrate
the surface tension because of their violent move-
ments and sink rapidly, dying by asphyxia on the
(2) Winged insects also show two types of be-
haviour. Those with some capacity to escape, es-
pecially hymenopterans such as winged ants and
wasps; dipterans such as house £ies and horse
£ies; snake£ies and lacewings; beetles such as
weevils, ladybirds and carabids; and true bugs
such as acantosomatids, pentatomids, pirrocorids
and reduviids fold the wings and raise them above
the water while trying to take o¡. If the insect
fails to £y or escape to the edge it dies exhausted
with the wings extended. Wasps £ex the abdomen
anteriorly when they die. The death throes do not
exceed 2 h. Cockroaches are unusually resilient ;
they can move rapidly and survive on the surface
of the water for up to 36 h. If the cockroach is
removed from the water during this time, it will
remain still for some minutes before £ying and
walking normally. Nevertheless, cockroaches are
common in carbonates from the Late Carbonifer-
ous to the Cretaceous. Their reduced abundance
thereafter is unlikely to be a re£ection of preser-
vation potential; the tegmina of blattoids have
been sclerotised since the Carboniferous. They
are rare in Tertiary limestones, although an ex-
ception is their elevated representation in the Pa-
laeocene of Menat, France (Nel, pers. commun.).
Cockroaches are more common in Tertiary than
in Cretaceous amber, where they are rare, except
for Lebanese amber (Azar, pers. commun.).
Winged insects that are unable to escape in-
clude small-bodied taxa with a wing span less
than 5 mm, such as planthoppers, aphids, £ies
and wasps and those taxa with a large wingspan
like butter£ies and grasshoppers. These taxa lack
su⁄cient strength to escape the water surface.
They become exhausted after a few minutes and
die by asphyxia within an hour. Winged grasshop-
pers remain lying on their side with the wings
folded dorsally. They can move on the water sur-
face but may die in this position. Laterally £at-
tened crickets are very common in limestones
's and Martinell, 1993 ; Pen‹alver,
1998). The death throes of insects produce surface
ripples that attract predators, including other in-
sects, ¢sh, birds and pterosaurs. Fish and croco-
diles usually eat the whole insect but wings and
highly sclerotised parts may be regurgitated intact
or defecated (Fig. 2B).
When aquatic insects die from environmental
agents or predation, the carcass does not have
to penetrate the surface tension to sink to the
bottom. For example, cases of caddis£y larvae
are abandoned as the insect emerges and may be
buried in the sediment, forming, for example, In-
Fig. 3. (A) Successive £ows (white arrows) that include several insects. EPGM-RD-0049; Lower Miocene, Hispaniola ; scale
bar = 2 mm. (B) Homoptera, Cicadellidae, covered by ¢ve successive £ows of resin. Insect was trapped in £ow 1 and ¢nally cov-
ered by £ow 5 (white arrows). EPGM-RD-0050; Miocene amber; Hispaniola, scale bar = 1 mm. (C) Swarm of ant imagoes.
MCAM0409; Miocene amber; Hispaniola, scale bar = 10 mm. (D) Isolated termite wings that separated along the humeral suture
after nuptial £ight. EPGM-CC-0051; Pleistocene^Holocene, Santander (Colombia); scale bar = 2 mm. (E) Platypodid borer beetle
swarm. MCAM0034; Miocene amber, Hispaniola; scale bar = 2 mm. (F) Small living insects trapped by surface tension and de-
composed before sinking, scale bar = 2 mm. (G) Butter£y wing scales (grey arrow) displaced from the body by successive £ows,
after partial decomposition; black arrow shows the £ow direction ; copal stalactite, Mart|
¤n-Closas collection, EPGM; Holocene,
Madagascar; scale bar = 2 mm.
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^64 29
dusia cypridis from the Late Eocene of Florissant,
Colorado, a case constructed of the shells of the
ostracod Cypris £orissantensis (Meyer, 2003). Su-
katsheva (1982) even proposed a Mesozoic and
Tertiary continental biostratigraphy based on
caddis£y cases.
Insects may also be introduced into aquatic set-
tings following forest ¢res. Grimaldi et al. (2000)
found a large diversity of fusanised insect remains
in Upper Cretaceous clays from New Jersey that
also yielded fossiliferous amber. These fusanised
insects represent a fossil scrub community prone
to wild¢re. Jarzembowski (2001a) reported beetle
remains preserved as fusain from the Lower Cre-
taceous of the Weald, southern England, showing
ultrastructural details.
6.1.2. Necrolysis in resins
Amber, in contrast to sedimentary rock, may
retain evidence of the death throes of the insects
fossilised within it. For example, wing movement
(Fig. 2C), disarticulation of body parts (Fig. 1D^
E), escape of decay gases, typically as a stream of
bubbles (Fig. 2F), and egg laying (Fig. 2G), are
commonly preserved. Poinar (2000) reported the
preservation of spermatophores with sperm in a
springtail in Baltic amber. Shedding of spermato-
phores commonly occurs during decay of Recent
insects, but is rarely observed in the fossil record.
The oldest known example is from a tiny wasp
from the Lower Cretaceous A
Łlava amber of Spain
(Fig. 8C).
Resin may form in plant parenchyma cells and
accumulate in internal channels and cracks within
the wood and under and within the bark as well
as on the surface of the tree (Fig. 2H). Sometimes
trunks may even burst due to resin pressure.
When resin escapes it may form stalactites, drops
and £ows (Weitschat and Wichard, 1998). The
position of these structures a¡ects their exposure
to polymerising agents. Free radicals, which are
induced by solar radiation and exposure to air,
contribute to the polymerisation of Hymenaea res-
in (Cunningham et al., 1977). Polymerisation of
Agathis resin is also thought to be dependent on
exposure to both air and light (Whitmore, 1977).
Alternating £ow and drying of resin results in the
di¡erentiation between darker and lighter layers
(Fig. 3A). Resins are produced not only by bark
and wood but also by roots and leaves. Neither
variation in the position and type of resin, nor its
impact on insect inclusions, have yet been studied
in the amber record.
Trees may produce several di¡erent kinds of
resins with di¡erent functions. The antimicrobial,
antifungal resin of the kaur|
¤tree Agathis australis,
for example, produces ¢ve types, depending on
whether it is exuded by leaves, trunk, branches
or roots (Henwood, 1992a; Langenheim, 1995).
The amount of secretion is controlled by the vis-
cosity of resin, which is determined by both envi-
ronmental temperature and the internal pressure
of the sap. Both temperature and pressure are
higher during spring and summer, with the result
that resins are exuded more frequently. Drops and
£ows form on the bark, usually daily, stalactites
on bark and branches, usually yearly. During
higher daily temperatures, resin is £uidised and
exuded by the tree, while during the night, the
temperature decreases, causing an increase in vis-
cosity and slowing down of exudation. Resin pro-
duction also is known to £uctuate seasonally.
This has been con¢rmed in Baltic amber based
on the presence of the stellate hairs of oak tree
£owers which appear predominantly during
spring and early summer (Weitschat and Wichard,
1998). Richardson et al. (1989) recorded that up
to 60% of the Baltic amber contains these hairs,
and was therefore produced at the same time as
the major seasonal abundance of insects. Sontag
(2001) surveyed 3875 pieces of amber (42.6 kg), a
random sample from unselected Baltic material,
and found that only 28% contained these plant
remains. Similar stellate hairs occur in Lebanese
amber, and may be a product of other plants
(Azar, pers. commun.). In any event, seasonal
changes in resin production in£uence the fossil
assemblage preserved in copal and amber.
A number of explanations have been o¡ered for
resin production: (1) as a defense mechanism
against fungal infestation or insect attack (True
and Snow, 1949; Janzen, 1975; Farrell et al.,
1991); (2) as a reaction to physical damage
(Meyer and Leney, 1968; Henwood, 1992a) ; (3)
as storage of unwanted products of cellular me-
tabolism or growth (Henwood, 1993) ; (4) as a
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^6430
protective barrier to reduce temperature and
water loss (Dell and McComb, 1978 ; Langen-
heim, 1995); and (5) to attract insect pollinators
(Langenheim, 1994; Goncalves-Alvim, 2001). The
¢rst two of these functions are thought to be the
most important. It is well documented that some
resin components act as herbivore deterrents
(although they may also act as attractants ; see
Labandeira et al., 2001) and have fungicidal prop-
erties (Messer, 1985; Richardson et al., 1989).
Some chemical compounds (such as alkanes and
monoterpenes) in conifer resins are toxic to fungi
transmitted by boring beetles (Cobb et al., 1968).
Some resins have insecticidal properties, e.g. resin
from Dipterocarpus kerrii kills workers of the ter-
mite Zootermopsis angusticollis (Richardson et al.,
1989). However, copious production of resin
could also be a reaction to damage. The e¡ects
of hurricanes may extend for hundreds of square
kilometres (Vandermeer et al., 2000), and living
Hymenaea courbaril is prone to hurricane damage
(Wadsworth and Englerth, 1959). It is likely that
the Caribbean was subject to tropical storms in
the Tertiary, just as it is today. However, obser-
vations on Hymenaea courbaril in Panama
showed that the controls on copious exudation
are complex, and physical damage was not the
only cause (Henwood, 1992a, 1993).
Saproxilic beetles depend on the availability of
damaged wood. When the woody structure of a
tree is damaged, it normally creates a chemical or
physical barrier with phytolexins to prevent inva-
sion by insects (Lehong and Laks, 1994). Oleor-
esins and other resins provide such a barrier and
usually trap saproxilic organisms. Interaction be-
tween the tree and beetles takes place during the
earliest stages following damage (Mamaev, 1961).
For example, the Buprestidae, a family of beetles
whose larvae often rely exclusively on wood, bore
large tunnels that also permit other organisms
to invade the wood. Similar wood-boring beetles
are ¢rst recorded in the Triassic, and other xylo-
phagous families such as Cerambycidae and sa-
proxilic ones such as Elateridae and Eucineti-
dae appear during the Jurassic (see Ross and
Jarzembowski, 1993; Labandeira, 1994). Never-
theless, the ¢rst record of substantial amber pro-
duction and of insect inclusions in amber is Early
Cretaceous. Molino-Olmedo (1999) showed that
78% of beetles embedded in amber are associated
with a saproxilic habitat, while 64% of saproxilic
families of beetles are ¢rst recorded in the Creta-
ceous. During the Early Cretaceous 32% of the
total number of saproxilic families of beetles ap-
pear. The ¢rst record of the beetle family Scoly-
tidae is also Lower Cretaceous ; this family in-
cludes important borers in the Oligo^Miocene
Hymenaea forests that produced Dominican and
Mexican ambers and also in cool-temperate
swamp forests near the North Pole that contained
resin-producing conifers (Labandeira et al., 2001).
26.6% of all families of beetles appear in the Low-
er Cretaceous, and 73.7% of these are related to
saproxilic habitats; all constituent species found
in amber belong to these saproxilic families (Mo-
lino-Olmedo, 1999). Bored wood from amber-
bearing lithologies from the Lower Cretaceous
of Archingeay, France, may have been invaded
by xylophagous beetles such as Cerambycidae,
Buprestidae and Scolytidae, and termites (Ne
deau et al., 2002). The ¢rst records of termite
sociality are also from Lower Cretaceous amber,
namely a few specimens from Jezzine and Ham-
mana, Lebanon (Azar, 2000), and many from
Montsec, Spain (Mart|
's and Martinell,
1995). In the Lebanese amber a few adults of dry
wood termites (Kalotermitidae), which live in ill-
de¢ned galleries inside the wood on which they
feed, have been found. In Spain many specimens
of adults and workers of damp-wood termites
(Hodotermitidae), have been found ; they also
feed on dead wood and may live in underground
colonies, although they may invade the tree di-
rectly without ground contact.
The evolution of herbivorous insects can be
compared with the phylogeny of resin producers
among angiosperms and gymnosperms (Dussourd
and Eisner, 1987). The fossil record of amber
shows that resin production became widespread
during the Early Cretaceous, coincident with the
expansion of some groups of boring insects (Chal-
oner et al., 1991). It has been suggested (Molino-
Olmedo, 1999) that this expansion of araucaria-
ceans and other resin-producing conifers coincides
with the ¢rst occurrence of some groups of xy-
lophagous insects, principally beetles and ter-
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^64 31
mites. Although there may be a link between the
Early Cretaceous radiation of araucariaceans and
the appearance of new groups of xylophagous in-
sects, this remains to be documented (but see Se-
queira and Farrell, 2001); the xylophagous insect
record extends back to the Triassic (Labandeira,
Palaeogeography determines not only the distri-
bution of the resin producers, but also what in-
sects are available for trapping. Henwood (1993)
argued that periods of extensive production do
not re£ect pathological conditions; the quantity
of resin produced by healthy modern trees is suf-
¢cient to explain the amount of amber in the fos-
sil record. Alternative explanations, not all of
which are amenable to testing with fossil deposits,
include: (1) prolonged and regular storm damage
due to palaeoenvironmental changes; (2) abun-
dance of mature trees, as resin production is low
in immature trees (Henwood, 1992a); or (3) high
prevalence of forest ¢res that lead to bark split-
An important question is why so much amber
does not include fossil insects. Only ¢ve of 65
Lower Cretaceous amber localities in Lebanon
yielded fossil insects (Azar, 2000). Similarly, of
more than 45 Lower Cretaceous amber localities
in Spain, only ¢ve yielded organism inclusions,
and in only one, A
Łlava amber of Pen‹acerrada, is
the number of individuals abundant ( s4000
specimens). It is unlikely that resins that do not
contain insects formed within the bark of trees,
because amber formed in such situations should
display a platy or lenticular morphology. Some
amber without insect inclusions may have accu-
mulated in soil, but only occasional pieces contain
a high quantity of organic matter or sediment,
and the soil organisms that might be trapped
are usually not present. However, many soil or-
ganisms are present, for example, in the Lower
Cretaceous amber of Archingeay, France (Ne
deau et al., 2002). Another hypothesis is that the
large quantity of resin without inclusions may
have been exuded into water, perhaps in response
to a £ooding event. While this is unlikely to ex-
plain the volume of resin that lacks insects, it
could be tested by observations of extant resin-
producing trees. The concentration of amber in-
clusions in a relatively small number of pieces
from a restricted number of outcrops remains to
be explained.
Resin trapping may be biased toward certain
organisms (Brues, 1933; Poinar, 1992; Henwood,
1992a). A number of factors promote preserva-
tion of some groups of insects and not others :
(1) Resin viscosity. Resin viscosity, stickiness,
and drying depend on its volatile content, which
controls the e¡ectiveness and longevity of the res-
in as a trap. These properties of the resin are
in£uenced by its position on the tree. The more
viscous the resin the higher the surface tension,
and the less the likelihood that it will be pene-
trated by insects. Large insects can penetrate
more easily than smaller ones, but they are also
more likely to struggle free (Henwood, 1993);
there are few records of large insects and small
vertebrates in amber. Large insects may, however,
adhere to the resin surface and be predated (Fig.
3B) before a new £ow covers them completely, or
they may leave appendages behind (Weitschat and
Wichard, 1998). Amber inclusions tend to be
dominated by small organisms. An investigation
of insect size revealed a predominance of small to
very small individuals of less than 5 mm long in
Cretaceous ambers while Tertiary ambers include
not only small insects but also middle to large
sized ones of more than 5 mm long (Poinar,
1992). However, these di¡erences may be the re-
sult of variation in the viscosity of the resin or
durability of the amber. The intense polymerisa-
tion of Cretaceous ambers renders them extremely
fragile and large insect specimens are rare (Azar,
2000; Grimaldi et al., 2000): less than 3% of more
than 2500 insects prepared from the A
Łlava Lower
Cretaceous amber are more than 4 mm long
's et al., 1999; Alonso et al.,
(2) Insect behaviour. The behaviour of insects
in£uences the likelihood of their entrapment in
amber. Certain categories are more at risk than
others: insects that live in bark or usually rest on
it, xylophagous insects that bore into wood and
bark or feed on leaves (Poinar and Poinar, 1999),
and swarming insects including dipterans such as
Chironomidae and Dolichopodidae and hyme-
nopterans such as Formicidae (Fig. 3C). Swarm-
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^6432
ing favours the inclusion of large numbers of in-
dividuals in amber (Koteja, 1996). Isolated wings
of adult termites are preserved in Pleistocene^Ho-
locene copal from Colombia (Fig. 3D) ; they shed
their wings after nuptial £ight by disassociation
along the humeral suture. Gregariousness pro-
motes the inclusion of large numbers of insects
in resin when it penetrates colonies of ants, ter-
mites and beetles (Fig. 3E). Up to 40 aphid speci-
mens are known from a single piece of Upper
Cretaceous amber from Grassy Lake, in Alberta,
Canada (Pike, 1993).
Some recent bees forage for resin (Henwood,
1992a; Goncalves-Alvim, 2001). This behaviour
increases the chance of becoming trapped and ex-
plains why Proplebeia is the most common bee
found in Dominican amber. Several specimens
of Proplebeia dominicana, with hind-leg baskets
¢lled with pollen and resin, have been reported
from Dominican amber (Roubik, 1989).
Some groups of parasitic and symbiotic insects
are brought to the trees by other organisms, or
carried by wind due to their small size. More than
90% of V400 hymenopterans in the Lower Creta-
ceous amber of A
Łlava, Spain, are small wasps that
parasitise other groups of insects (Mart|
's et al., 1999). Di¡erent larval stages of insects
parasitic on mammals or birds, such as haema-
tophagous adults, eggs or larvae, often make their
¢rst appearance in the fossil record in Eocene
Baltic amber (Larsson, 1978; von Schlee, 1990).
Phoretic arthropods, particularly mites, are usu-
ally entombed on the host insect (Larsson, 1978 ;
Grimaldi, 1996; Poinar et al., 1998).
Even aquatic insects may be trapped in resin.
Larvae of may£ies, caddis£ies and stone£ies, pu-
pae of some other groups of £ies, and adult ripar-
ian and limnic beetles and bugs are represented in
Tertiary ambers (Baltic: Weitschat and Wichard,
1998; Dominican: von Schlee, 1990; Poinar and
Poinar, 1999). Wichard and Weitschat (2001) re-
ported that at least 30% of Baltic amber inclu-
sions are aquatic insects, but this ¢gure includes
pupae, larvae and adults, not all of which are
aquatic. Aquatic insects represent 14% of all in-
clusions in the New Jersey Upper Cretaceous am-
ber (Grimaldi et al., 2000). Weitschat (2001) re-
ported the presence of adults of the ostracod
Cyclocypris in Baltic amber. Thus in some cases
there was presumably standing water near the am-
ber trees (Poinar, 1992; Poinar and Poinar, 1999;
Grimaldi et al., 2000).
Grimaldi (1996) recorded a crab from Miocene
amber from Chiapas, Mexico, which occurs in
lignites interbedded with marine sandstones and
silts. Associated pollen indicates that the amber
was deposited in a mangrove environment. Such
crabs live today in forests close to the seashore in
Costa Rica, in tunnels between the roots of Hy-
menaea courbaril (pers. observation). It is possible
that resin could have been eroded from the soil to
accumulate among the mangrove roots (Langen-
heim, 1995). Resins could also have accumulated
in tunnels and trapped the crab. Nevertheless, the
resin producer must have been rooted in water to
account for trapping of the crab ; transported res-
in cannot trap organisms.
(3) Insect habitat.Zherikhin and Sukatcheva
(1990) documented insect assemblages in resins
produced by living Eurasian temperate conifers
and showed that they contain a large number of
snake£ies, alder£ies, earwigs, and stone£ies. The
small number of £y species and the absence of
may£ies and caddis£ies emphasise that the com-
position of these assemblages is clearly distinct
from those in any known amber. This di¡erence
is attributed to the formation of known fossil
amber in tropical rainforests rather than temper-
ate conifer stands.
The location of resin production a¡ects the
trapping of insects. Insects that live around res-
in-producing trees are most prone to entomb-
ment. Small insects from elsewhere may have
been carried to the resin by wind, although this
might be prevented in very dense forest. This hy-
pothesis could be tested in New Zealand today in
forests of the araucariacean Agathis. Insects from
wet soil, bark, and leaf habitats are well repre-
sented in amber but those from habitats other
than forest, for example prairie or dry environ-
ments, are rarely represented (Krzeminska et al.,
1992; Poinar and Poinar, 1999). Accumulations
of Agathis resin persist in surface layers of soil
for thousands of years (Thomas, 1969). Resin
concentrations also occur around Hymenaea indi-
viduals where they not only fall from the tree, but
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^64 33
are produced by the roots (Langenheim, 1967).
Henwood (1992a, 1993) suggested that extensive
underground accumulations of resin occur prin-
cipally around the root systems of large trees.
Accumulations of resin are found buried around
Recent Agathis,Copaifera, and Hymenaea, some-
times up to 1.8 m below ground (Howes, 1949 ;
Whitmore, 1980; Langenheim, 1967). Some au-
thors (Howes, 1949) discussed whether such resin
and copal accumulations were produced total-
ly by roots or are the result of both aerial and
underground production. The presence of an
underground soil fauna in the amber suggests
that at least some of these resins were produced
in situ by roots. Amber has only been found
in association with a developed root system in
the Lower Cretaceous of Jordan (Nissenbaum
and Horowitz, 1992). Bandel et al. (1979) sug-
gested that the resin was concentrated in ‘par-
autochthonous’ lignite beds; the resins and other
tree remains that produced the lignite beds may
have been trapped by mangrove roots, although
such root systems are not characteristic of the
resin producers.
(4) Plant defence. Volatile resin products may
deter or attract groups of insects. Resins protect
trees from insects, fungi and bacteria. Floral res-
ins may attract pollinators to £owers (Armbrus-
ter, 1984). Some leaf-cutter ants are deterred by a
terpenoid compound secreted by the leaves of Hy-
menaea courbaril (Ales et al., 1981). In Recent
rainforests, diverse associations of resin-produc-
ing plants may enhance their defence against foli-
vores and pathogens (Langenheim, 1995). Pollen
of diverse resin producers, including araucaria-
ceans, cupressaceans and cheirolepidaceans, oc-
curs in the Lower Cretaceous amber of Spain
¤n et al., 2001). The diversity of such asso-
ciations is lower in more xeric environments.
(5) Environmental factors that in£uence resin
production. Abiotic factors, such as light, temper-
ature, moisture, and nutrients play an important
role in controlling the nature and quantity of res-
ins, and of insect inclusions (Langenheim, 1995).
Elevated temperature promotes resin exudation :
trees exude more resin during spring and summer
than in autumn and winter. For this reason the
majority of insects in Tertiary ambers are spring
and summer taxa. The unusual occurrence of rep-
resentatives of some living winter and crepuscular
dipterans, such as Trichoceridae and Culicidae,
respectively, in Baltic amber suggests that the res-
in-trapped insects lived proximal to the trees
(Krzeminska et al., 1992). Resin that is produced
underground is subjected to di¡erent and more
constant temperature and humidity than that in
the open air, and is protected from UV light,
which promotes polymerisation.
Soil type and water availability also in£uence
resin production. In the extant Copaifera mitijuga
forests of Brazil, trees living in clayey soils pro-
duce more resin than those living in sandy soils
(Alencar, 1982). Resin production is greater in
Hymenaea courbaril trees of similar size with
more available water (Langenheim, 1967). Solar
radiation also increases resin production in Hyme-
naea, whereas darkness and soil nitrogen reduce
it. Resin production may vary widely among in-
dividuals of the same species, even when they
occupy similar habitats (Adams, 1977). Addition-
ally, resin composition di¡ers between geograph-
ically separate populations of Hymenaea courbaril
(Langenheim, 1984).
Resin may be produced following splitting of
bark as a result of ¢res (Hillis, 1987). Siegert et
al. (2001) investigated exceptional ¢res in the dip-
terocarp-dominated rain forests of Borneo (Indo-
nesia) resulting from droughts promoted by ‘El
Nin‹o’ between 1997 and 1998. Dipterocarps are
angiosperms that are exploited to produce dam-
mar resin (Heywood, 1998). 2.6 million hectares
of rain forest were burned, resulting in damage
varying from moderate (25^50% trees killed) to
near total ( s80% vegetation killed). The ¢re af-
fected mainly lowland dipterocarps (40.5% of the
total area of distribution burned), as well as sec-
ondary forest and peat swamp forests. The high-
land dipterocarps were only partially a¡ected, by
68.5% of the total area of distribution burned.
Canarium strictum (Burseraceae) is also burnt to
induce the production of dammar resin, causing
tree exudations for about ten years (Henwood,
1993). Fusanised £owers and insects, and ¢re-
damaged amber, are found associated with creta-
ceous amber in New Jersey. These constituents
suggest that resin production may have been in-
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^6434
duced by ¢res (Grimaldi et al., 2000). Levels
of atmospheric oxygen during the Cretaceous
were among the highest during the Phanerozoic,
close to 30% compared to 20% today, leading to
a much higher prevalence of ¢res (Robinson,
(6) Other biological factors. Small disarticulated
fragments of insects and other arthropods are
common in amber. Kutscher and Koteja (2000)
o¡ered three explanations for the origin of these
inclusions. They are parts of exuviae, components
of the faeces of a predator, or detritus produced
by mites, microorganisms and/or physical factors.
They concluded that the inclusions are most likely
to be the remains of faeces. Insect faeces or frass
also have been recorded in amber (Grimaldi et al.,
2000; Ne
¤raudeau et al., 2002) as have prey car-
cass refuse heaps that have been attributed to spi-
ders (Weitschat and Wichard, 1998).
6.2. Biostratinomy
The ¢rst studies on insect biostratinomy were
published by Scha
«fer (1972) and Lutz (1984,
1990). Scha
«fer (1972) described the necrolysis
and biostratinomy of Recent Libellula quadrima-
culata, and other insect orders such as beetles,
£ies and wasps, from the North Sea. Lutz
(1984) described the disarticulation of one element
of cockroach elytra. In 1990 he discussed the dif-
ferent stages of insect preservation evident in the
Messel oil-shale (Eocene, Germany), and docu-
mented and interpreted the ¢rst actualistic experi-
ments on the biostratinomy of insect carcasses
(Lutz, 1990). Wilson (1980, 1988) compared insect
assemblages from nearshore and o¡shore settings
in several Palaeogene lakes in North America,
and concluded that insects are well represented
in nearshore deposits, although they are more
fragmented in o¡shore settings. Dipterans, hyme-
nopterans and heteropterans are better repre-
sented than other groups in o¡shore zones,
while trichopterans and coleopterans are more
common in nearshore zones. Mart|
and Martinell (1993),Lutz (1997),(Rust, 1998),
Pen alver (2002) and Duncan et al. (2003) have
contributed actualistic studies on insect biostratin-
6.2.1. Biostratinomy in aquatic settings
Biostratinomic processes are a very important
control on insect preservation in carbonates.
Winged insects that fall into water are a¡ected
by diverse biostratinomic processes (Fig. 4). The
time spent £oating varies even within the same
insect taxon (Mart|
's and Martinell,
1993). Insects that die in aerial or terrestrial con-
ditions and are later introduced to water £oat for
longer than insects that drown. Insects may £oat
in the calm water of tanks for more than 6
months. Where ¢shes are present, they eat the
majority of insects within a few hours and they
are clearly an important taphonomic ¢lter in
lakes, rivers and lagoons.
Insects with a large wingspan, such as butter-
£ies and dragon£ies, and very small insects cannot
penetrate surface tension and usually decompose,
disarticulate and fragment on the water surface
(Fig. 3F). Butter£ies have scaly wings that soak
easily, and when they fall onto water, they are
trapped by the surface tension (Mart|
and Martinell, 1993). Two days after death the
body shows considerable decomposition, and
any biological or physical disturbance promotes
fragmentation and rapid decay. Many hundreds
of individuals may be preserved in exceptional
circumstances, as are lepidopterans in o¡shore
marine diatomites in the Palaeocene of Denmark
(Rust, 2000). Rust (1999) interpreted these lepi-
dopteran accumulations as a result of rapid pen-
etration of surface tension due to strong wind or
rain. Where insects with a large wingspan and/or
very small ones are well preserved elsewhere
(Leestmans, 1983), this is often due to the devel-
opment of cyanobacterial mats (Mart|
and Martinell, 1993; Etter and Kuhn, 2000).
Observations on outdoor tanks in Barcelona
during the spring, summer and autumn of 1990^
1991 revealed a ratio of small to large ( s5mm)
insects of more than 25:1 counted every two days
for two weeks. This proportion is reduced consid-
erably after a rainy or windy day, but fragments
and disarticulated parts are present on the bot-
tom. Large insects may remain on the water sur-
face for 4^14 days (Mart|
's and Marti-
nell, 1993). Water entering the tracheal system
may increase the weight of the carcass su⁄ciently
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^64 35
to allow it to sink. Wetting and sinking are has-
tened by decomposition. Sinking is also promoted
by the growth of fungi or algae on the insect
carcass (Mart|
's and Martinell, 1993),
both of which usually begin on the articulations
of the abdominal segments. However, fungi and
algae may also hold the carcass together, inhibit-
ing disarticulation. Fungi and algae are preserved
through mineralisation in association with insects
from the Tertiary of Riversleigh, Queensland
(Duncan et al., 1998).
Experimental observations in tanks show that
sinking of an insect carcass is usually slow and
vertical. Orientation varies depending on the po-
sition in which the carcass is stable on the sub-
strate. This is true even for insects enveloped by
algae or fungi. Normally only insects trapped by
cyanobacterial mats in the water column exhibit
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^6436
varied positions. The presence of a thermocline,
with or without a halocline, represents a density
change that prevents sinking and promotes de-
composition before carcasses are deposited on
the bottom. Carcasses may remain articulated
for almost one year on the bottom of tanks in
undisturbed ambient conditions. Disarticulation
and fragmentation are promoted by biological
or physical agents (Tischlinger, 2001 ; Duncan et
al., 2003).
Carbonate rocks which yield fossil insects often
show evidence of anoxic conditions (Martill,
1993; Etter and Kuhn, 2000) and the insects
may be relatively complete. Disarticulation and
fragmentation presumably occurred during £oat-
ing and sinking in the oxygenated zone above a
thermocline or halocline, or as a result of dis-
placement by decay gases. Chitinophagous bacte-
ria that degrade cuticle live almost exclusively in
oxygenated environments, although some are ca-
pable of anaerobic metabolism.
The disarticulated elements that commonly oc-
cur in carbonate rocks are: head and thorax, head
with antenna, thorax with legs, thorax with wings,
and thorax and abdomen. Other isolated struc-
tures, including palps, cerci and tarsi, are rare.
Early disarticulation and disintegration may be
the result of predation and scavenging by verte-
brates and invertebrates at the air^water interface,
sediment^water interface, and within the sedi-
ment. It may be di⁄cult to distinguish fragments
left by predators from those produced by physical
agents. Predators and scavengers may control the
composition of fossil arthropod assemblages
(Plotnick, 1986, 1990). Their absence is critical
for the preservation of insect carcasses (Duncan
et al., 2003). Fragments of insects also occur as
gut contents of amphibians (Pen alver, 2002),
¢shes (Richter and Baszio, 2001), some mammals,
such as bats or anteaters (Schaal and Ziegler,
1992), and in coprolites (Ansorge, 1993)(Fig. 2B).
6.2.2. Biostratinomy in resins
The biostratinomy of insects entombed in resins
can be regarded as commencing only when the
resin is ¢rst transported. By this de¢nition the
entombment of the insect in the resin is not syn-
onymous with burial of a carcass in sediment, but
is a pre-burial process that generally takes a very
short time, from seconds to a few days. However,
it is not always possible to di¡erentiate between
factors a¡ecting insect inclusions before and after
resin burial, and therefore the process of entrap-
ment is also considered here.
Insects usually die by asphyxia and only large
insects, which require more than one resin £ow to
overwhelm them, perish as result of fatigue or
predation. The time required for resin to over-
whelm an insect depends on its viscosity, the fre-
quency of £ow, and how the resin is deposited : if
resin falls as drops (Wagensberg et al., 1996),
burial is rapid. When insects are trapped in resin
by the legs or wings, another drop may disartic-
ulate the carcass, particularly when it has dried
(Fig. 3G). Twisted wings, abnormal positions, or
assemblages displaying oriented body structures
(legs, antennae, scales, etc.) usually re£ect resin
£ow (Fig. 5A).
Occasionally amber preserves alternating clear
and dark bands, representing successive £ows
Fig. 4. Carbonate taphonomy. (A) Travertines form in £uvial settings or springs with high alkalinity, and preserve terrestrial and
aquatic insects. (B)^(D) Carbonate rocks preserve terrestrial, semiaquatic and aquatic insects. Unlike aquatic insects, terrestrial
and semiaquatic taxa need to penetrate the surface tension before sinking. (E) During sinking insects may be trapped by micro-
bial mats that live in the water column, or they may be coated by algae or fungi that inhibit disarticulation. Microbial mats also
grow at the sediment^water interface; where they also trap insect carcasses. (F) Sinking may be delayed by the thermocline or
halocline; anoxic bottom conditions delay decomposition. (G) Travertines may be produced under water, but insects are rarely
preserved; typically plant and algal remains are found. (H) Immediately after burial, insects may be mineralised by diagenetic
minerals such as calcite or pyrite. (I) Insects may decay, or the early diagenetic minerals may dissolve to leave a cavity in which
other minerals may precipitate, forming a cast. (J) When pressure and temperature a¡ect fossil insects, the volatile components
of organic structures may be lost and the original organic compounds diagenetically transformed. (K) In travertines organic mat-
ter is rarely preserved, and insects and plants are usually found as moulds. (L) The space left by degradation of the insect may
or may not be in¢lled by a second generation of minerals. (M) Carbonate rocks that preserve fossil insects are usually those de-
posited in continental or shallow marine environments.
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^64 37
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^6438
(Weitschat and Wichard, 1998). The darker
layers, usually thinner than the clear ones, are
the result of rapid drying by sunlight and wind.
These dark layers may contain insects that were
not completely encapsulated until the arrival of
the next £ow or drop of resin (Fig. 3B). Some
examples show that insect structures, such as
wings, can prevent rapid drying of the resin
(Fig. 5B).
The majority of amber outcrops are allochtho-
nous. Only in the Lower Cretaceous of Jordan is
amber found in association with tree roots, possi-
bly the producer plant (Nissenbaum and Horo-
witz, 1992). Amber may occur in association
with coal seams, as in the Upper Cretaceous of
Grassy Lake, Canada, or in the Lower Cretaceous
of Wadi Zerka, north of Amman, Jordan (Poinar,
1992; Bandel et al., 1979). Some authors (Pike,
1993; Bandel et al., 1997) have interpreted such
deposits as ‘in situ’, but they may represent trans-
ported accumulations. Where fragile structures
such as stalactites or individual drops are pre-
served, it suggests that they have not been trans-
ported far, or that they have been carried by
water without any contact with the bottom sedi-
ment. The presence of aquatic animals in amber
indicates the proximity of water to the trees. Some
species of Hymenaea in the Amazonian rainforest
today may be periodically inundated (Langen-
heim, 1995), and likewise Agathis australis of
New Zealand.
The transport and deposition of resins are
poorly understood (Fig. 6). Amber is rarely found
within fossil wood, suggesting that they are nor-
mally transported separately or that the amber is
destroyed during diagenesis. Clastic sediments
that yield amber are frequently rich in organic
matter, but rarely in wood or bark fragments.
Where amber is preserved in wood, this may pro-
vide important evidence of the source plant and
the part of the tree from which the amber origi-
nated. Small pieces of amber occur in the wood of
Metasequoia glyptostroboides from the Middle
Eocene of Axel Heiberg Island, probably as a
result of insect infestation (Labandeira, pers. com-
mun.). Dejax in Azar (2000) suggested that Leb-
anese amber is produced by the cheirolepidacean
Protopodocarpoxylon sp., congeneric with P. sub-
rochii, from the Upper Jurassic of Morocco, and
not by an araucariacean as previously thought.
Ghiurca (1988) proposed a taxodiacean origin
for Romanian Miocene amber. Knowlton (1896)
proposed a cupressacean origin for some pieces of
amber from the Upper Cretaceous of New Jersey,
but Grimaldi et al. (2000) considered a pinacean
or taxodiacean origin more likely based on chem-
ical analysis of the amber.
Although resins are produced by plants that
would be expected to accumulate in deltaic and
lacustrine sediments, remains of resins are rarely
reported in studies of plant taphonomy in these
environments. Gastaldo et al. (1987) reported no
resin in their taphonomic study of plant remains
in the Holocene crevasse splay in the Mobile Del-
ta of Alabama, USA, nor in the backswamps of a
Holocene oxbow lake of the Alabama River,
where an important swamp cypress (Taxodium)
community lived (Gastaldo et al., 1989). When
resins occur in association with other plant re-
mains in some sedimentary environments, they
tend to be preserved as rare fragments. Resins
are found as palynodebris in most delta-plain sub-
environments associated with parenchyma tissues
and wood debris, and sometimes in distributary
channels, such as the Middle Miocene of the Ni-
ger Delta amid an angiosperm dominated forest
Fig. 5. (A) Termite, Rhinotermitidae, with rolled wings. EPGM-RD-0052; Miocene amber, Hispaniola ; scale bar = 2 mm. (B)
Wasp, Scelionidae, showing that wings inhibit rapid drying and polymerisation of the resin (dark layer). MCNA-9244; Lower
Cretaceous, Moraza (Spain); scale bar = 1 mm. (C) Ants preserved in large numbers due to precipitation of authigenic carbonate.
TJ-0213-MCCM; Miocene, Tres Juncos (Spain) ; scale bar = 2 mm. (D) Neuroptera, Chrysopidae, showing elongation of left
forewing due to tectonic deformation. ADR-0029-I/MCCM; Barremian, Las Hoyas (Spain); scale bar = 5 mm. (E) Dicranomyia,
Diptera, Limoniidae, preserving delicate structures. MPV-69-RM; Lower Miocene, Rubielos de Mora (Spain); scale bar = 5 mm.
(F) Aktassoblatta fusca, cockroach preserving organic rich cuticle. PIN 2239/347; Upper Jurassic, Karatau (Kazakhstan); scale
bar = 5 mm.
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^64 39
(Oboh, 1992). Phytoclasts in the Recent Mahakan
River Delta in Borneo (Gastaldo and Huc, 1992)
include dipterocarpacean resins (dammar) as ei-
ther duct in¢lls or amorphous to cylindrical clasts
associated with other plant debris. These struc-
tures occurred in high energy depositional envi-
ronments. Resin clasts may be up to several cen-
timetres long and over 1 cm in diameter, and
usually are rounded. These resins account for up
to 4.5% of plant parts recovered on the delta front
and in tidally in£uenced interdistributary areas
(Gastaldo, 1994). The Lower Cretaceous amber
of A
Łlava, Spain, occurs in sandy channels of the
delta plain (Alonso et al., 2000).
6.3. Diagenesis
Insect carcasses that survive biostratinomic pro-
cesses may be buried (Fig. 4). Subsequently they
may be a¡ected by a number of processes: (1)
early mineralisation; (2) £attening; (3) deforma-
tion; (4) thermal maturation ; and (5) reworking.
Other factors, such as bioturbation, decomposi-
tion and weathering also may lead to biases in
fossil insect assemblages.
(1) Early mineralisation. The preservational
state of fossil insects depends on the degree of
decomposition when an insect arrives at the burial
environment. Decay-resistant tissues may survive
to be preserved as organic material ; diagenetic
changes result in transformation to more stable
molecules than those of the original (Briggs,
1999). Preservation of labile tissues requires rep-
lication by authigenic minerals before the loss of
morphological detail. The type of mineral in-
volved depends on the chemistry of the environ-
ment and the prevalent microbial processes (Efre-
mov, 1950; Allison, 1988). Both types of
preservation ^ of resistant and labile tissues ^
may occur at the same fossil locality and even
within the same fossil (Penalver et al., 1993).
(2) Flattening. Flattening includes collapse of
the carcass due to decomposition, and compac-
tion as a result of overburden pressure. The de-
gree and nature of £attening is determined by the
grain size and composition of the sediment, the
morphology and composition of the carcass, its
orientation to the bedding, the nature and timing
of diagenesis, and processes involved in cavity in-
¢ll. 3-D preservation of insects in limestone im-
plies that mineralisation of the soft-tissues (Dun-
can and Briggs, 1996) and/or early cementation of
the surrounding sediment (McCobb et al., 1998)
was initiated prior to appreciable decay (Fig. 5C).
The e¡ects of compaction may be di¡erentiated
from tissue collapse through evidence of distor-
tion of the mineral in¢ll, replacement, or coating
(Allison and Briggs, 1991a). 3-D preservation is
normal in amber. Subsequent compaction of the
host sediment may result in ‘engraving’ of the
surface and £attening of a piece of amber (Alonso
et al., 2000; Zherikhin and Ross, 2000) but rarely
a¡ects the inclusions.
(3) Deformation. Fossil insects are sometimes
distorted tectonically and this may lead to taxo-
nomic misidenti¢cation (Pen alver, 1996)(Fig.
5D). Limestone is more resistant to weathering
and erosion than amber, which may be very frag-
ile due to intense polymerisation.
(4) Thermal maturation. Maturation results in
the transformation of the organic components of
a fossil insect to a kerogen-like composition
(Stankiewicz et al., 2000). It causes amber to be-
come darker and more fragile by promoting the
loss of volatile terpenes and increasing the rate of
polymerisation. It may also increase amber plas-
(5) Reworking. Amber with insect inclusions is
often reworked. Fossil insects in carbonates are
only likely to be reworked if they are preserved
in nodules.
6.3.1. Diagenesis in carbonate rocks
Rapid burial favours exceptional preservation
(Seilacher et al., 1985) by eliminating scavengers
and promoting anaerobic conditions that inhibit
bioturbators. It results in the preservation of ar-
ticulated carcasses (Fig. 5E). Low rates of decay
are necessary to allow authigenic minerals to form
quickly enough to preserve the most labile soft-
tissues (Briggs, 1995b). If decay outpaces miner-
alisation, only the most decay-resistant tissues and
structures will be preserved. Organic remains sur-
vive only where chemical (hydrolysis, oxidation)
and biological (enzymatic and microbial activity)
degradation is prevented (Briggs, 1999).
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^6440
Fig. 6. Amber taphonomy. (A) Terrestrial, aquatic and underground insects are trapped by resin. (B) Resin may accumulate in
the internal cracks and pockets within the wood, under and between the bark. (C) When resin is unconstrained it may form sta-
lactites, drops and £ows, and may trap insects and other organisms. Under subaerial conditions resins lose volatiles. (D) Subter-
ranean deposits of resins also form, produced by roots and by the aerial parts of the tree, and accumulate as large masses
around the tree base. (E) In the majority of cases it is not known whether resin is transported to the deposit in which it becomes
fossilised with the tree or separately. (F) Resins are introduced into water directly from the tree or following erosion of the soil.
(G) Initial deposition of the resin, usually associated with organic rich sediments. (H) Diagenesis of the resin begins with burial.
Nevertheless, some prediagenetic processes that a¡ect insects in resins are di⁄cult to distinguish from the e¡ects of diagenesis
(see text). (I) Amber is usually reworked and deposits are time-averaged.
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^64 41
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^6442
Insects are preserved in carbonate rocks in six
major ways: (1) as cuticle remains, usually as a
diagenetic product of the original organic compo-
sition; (2) by preservation in an authigenic min-
eral, often calcium carbonate, pyrite or calcium
phosphate, or in¢lling of a decay void by minerals
such as calcite, silica, or pyrolusite which may be
recrystallised or altered during diagenesis ; (3) as a
mould of the exoskeleton, resulting from the loss
of organic remains; (4) in travertine, usually as a
mould or carbonate coat; (5) in a concretion,
often as a moulds or through replication by au-
thigenic minerals. More than one type of preser-
vation may occur at the same locality and in the
same specimen.
(1) Cuticle remains. Insects are usually pre-
served as cuticle remains (Fig. 5F). Loss of the
organic remains may result in preservation as a
mould. Both of these types of preservation com-
monly occur in the same sequence : in the lami-
nated Lower Cretaceous limestones of Las Hoyas,
Spain, for example, cuticle commonly survives in
dark laminae, but has been lost in the lighter col-
oured laminae (Fregenal-Mart|
¤nez et al., 1992).
Cuticle may be preserved in three dimensions
due to sclerotisation or early mineralisation of
the enclosing lithology (McCobb et al., 1998).
Aquatic Heteroptera are preserved in three di-
mensions at Las Hoyas where the content of or-
ganic matter in the sedimentary matrix is low.
Where more organic matter is present, compac-
tion results in 2-D preservation (Mart|
et al., 1995).
The chemistry of the cuticle of fossil arthropods
has been investigated using pyrolysis^gas chroma-
tography/mass spectrometry which allows insolu-
ble macromolecular material to be analysed in
very small quantities (Stankiewicz et al., 1997a ;
Briggs et al., 1998 ; Briggs, 1999). Arthropod cu-
ticles are composed of chitin and proteins, cross-
linked via catecholamine and histidyl moieties
(Stankiewicz et al., 1997b). Traces of cuticle chitin
and protein may survive in Cenozoic strata : the
oldest so far reported are in a curculionid beetle
from Oligocene volcanoclastic lacustrine sedi-
ments at Enspel, Germany (Stankiewicz et al.,
1997b). In Eocene and older deposits, the organic
components of arthropod cuticle are altered to
aliphatic compounds, as in the water boatman
Iberonepa romerali from Las Hoyas in Spain
(Stankiewicz et al., 1997a). Aliphatic pyrolysates
are interpreted as a product of polymerisation of
free or ester-bound aliphatic compounds and the
simultaneous alteration of labile polymers such as
proteins and chitin (Stankiewicz et al., 2000).
Such a process may also lead to the formation
of aromatic moieties (Harvey et al., 1983) in fossil
crustaceans and other organisms (Stankiewicz et
al., 1997a).
(2) Mineralisation. There are three main modes
of mineralisation of insects: preservation of tissue
morphology, pseudomorphing and in¢lling to
form a cast. Labile soft-tissues may be replicated
in apatite (Duncan et al., 1998), which preserves
remarkable detail, or in pyrite/goethite. Decay-re-
sistant cuticles may be fossilised by the precipita-
tion of authigenic minerals such as apatite or py-
rite in their structural spaces, as in the insects of
the Crato Formation in Brazil (Grimaldi and
Maisey, 1990). Labile soft-tissues are more com-
monly preserved in outline, by bacteria that colo-
nise the carcass and become mineralised them-
Fig. 7. (A) Cymatophlebia longialata, dragon£y, with wing venation replaced by pyrolusite during diagenesis. 1959.73.Ka/J.M.E.;
Upper Jurassic, Solnhofen (Germany); scale bar = 10 mm. (B) Fly preserved in silica in a nodule from the Barstow Formation;
Middle Miocene, Southern California (USA), photo by L. Park; scale bar = 1 mm. (C) Aquatic Heteroptera preserved secondarily
in goethite, originally in pyrite, EPGM-RD-0053; Lower Cretaceous, Crato Formation (Brazil) ; scale bar = 10 mm. (D) Abdomi-
nal apex of caddis£y larva preserved in authigenic calcium phosphate, QM F34594, Oligocene^Early Miocene of Riversleigh
(Queensland), scale bar = 0.1 mm. (E) Beetle 3-D preserved in authigenic calcium phosphate, QM F34595; Oligocene^Early Mio-
cene, Riversleigh (Queensland) ; scale bar = 0.5 mm. (F) Reisia sodgianus, isolated dragon£y wings preserved as moulds. PIN
2785/4; Upper Triassic, Dzhayloutcho (Kirghizistan); scale bar = 5 mm. (G) Aeshna isosceles, dragon£y hindwing preserved as a
mould in 5000-year old travertine. IPM-R.08356; Holocene, Auriol (France), photo by A. Nel ; scale bar = 5 mm. (H) Carbonate
nodule from Lower Miocene, Izarra (Spain), IZA-001; scale bar = 10 mm. (I) Imago ant preserved as a mould in the Izarra nod-
ule, JFP-001 ; scale bar = 5 mm.
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^64 43
selves. Such microbial ¢lms can be preserved in
various minerals. This type of preservation has
been recorded in animal tissues preserved in the
deposits of Eocene Lake Messel (Wuttke, 1983),
for example, but has not been reported in the
insects from the same site, which are preserved
as diagenetically altered organic material. In¢lling
occurs where the void left following decay is sub-
sequently ¢lled by a mineral, as in pyrite in¢lled
beetles from the Eocene London Clay of south-
eastern England (Britton, 1960; Jarzembowski,
1992). The diagenetic minerals that most com-
monly form casts are calcium carbonate, pyrolu-
site and silica (Fig. 6A,B). In the Cretaceous of
Las Hoyas and the Jurassic Solnhofen Limestone
of Bavaria, calcite and pyrolusite precipitate in
the spaces left following decay and form casts.
Sometimes pyrolusite is con¢ned to the wing
veins. In contrast, calcite in¢lls the void left
when insects decay in the Eocene Bembridge
Limestone and may replicate the soft-tissues
(McCobb et al., 1998). Silica, celestine, apatite
and gypsum form casts and replace the cuticle
of the insects in carbonate nodules from the Mio-
cene Barstow Formation, USA (Park and Down-
ing, 2001).
A variety of authigenic minerals may replace
soft-tissues (Allison, 1988). Labile tissues can
only be replicated where mineralisation is rapid
relative to decay. The principal authigenic miner-
als that replace insect soft-tissues in carbonate
rocks are calcite/aragonite, apatite and pyrite
(Fig. 7C). In most freshwater and marine environ-
ments, concentrations of the bicarbonate ion ex-
ceed those of the phosphate ion, and calcite/ara-
gonite formation is favoured. The solubility
product of calcium phosphate must be greater in
order to precipitate apatite (Briggs and Wilby,
1996). Early diagenetic pyritisation requires rapid
burial, low organic matter content, and the pres-
ence of sulphates, and it is favoured by high con-
centrations of reactive iron (Brett and Baird,
1986; Briggs et al., 1996).
Decay can be rapid under anaerobic conditions.
Preservation of soft-tissues in such circumstances
is due mainly to the link between decay and
mineral formation, rather than to limited decom-
position. Mineralisation is favoured by elevated
microbial activity, which establishes the geochem-
ical gradients required for precipitation. The locus
of mineral precipitation is determined by geo-
chemical gradients from the decaying organism
into the surrounding sediment (Can¢eld and Rais-
well, 1991).
Calcium carbonate and calcium phosphate
(apatite) are the two major types of mineral that
occur in limestones. The controls on the forma-
tion of these minerals are amenable to experimen-
tation (Briggs and Kear, 1994; Sagemann et al.,
1999). For example, shrimps decayed in marine
media at 15‡C under oxic conditions underwent
limited weight loss when indigenous bacteria
alone were present. This resulted in the formation
of a variety of CaCO3crystal bundles, including
discs, rods and dumbbells, but there was no soft-
tissue preservation. In contrast, where decay rates
were elevated by the addition of an inoculum of
bacteria, anaerobic sulphate reduction was in-
tense, causing more extensive weight loss, pH de-
creased, and some muscle tissue was replicated in
apatite. Thus, while calcium carbonate is nor-
mally favoured due to the high concentrations
of the bicarbonate ion, it may be inhibited due
to reduced pH, allowing apatite to form. A limit-
ing factor is that continental water bodies are
characterised by low concentrations of phospho-
rus. Only highly alkaline waters, poor in Ca, re-
tain much phosphate in solution. Consequently,
calcium carbonate minerals are normally associ-
ated with fossil insects. Bacterially induced
CaCO3precipitation results in single crystals or
in crystal bundles in the form of rods, spheres,
and brushes (Buczynski and Chafetz, 1991). In a
liquid medium, circulation and ion di¡usion rate
are both high, precipitation is rapid, and arago-
nite forms. Where di¡usion is lower, as within the
gelatinous secretions of bacteria and/or algae, cal-
cite is precipitated.
Calcium carbonate does not normally replicate
the structure of soft-tissues (but see McCobb et
al., 1998). Its importance in preserving insects is
in in¢lling voids left following the decay of soft-
tissues, but more importantly in promoting rapid
lithi¢cation, particularly the formation of nod-
ules. In the eutrophic conditions of some lakes
and lagoons, high productivity of organic matter
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^6444
is associated with anaerobic bottom conditions.
Bicarbonate ions released by decay react with cal-
cium or iron to produce authigenic carbonates.
Limestone precipitation is favoured by bicarbon-
ate saturation that results from CO2production
by bacteria, algae, and plankton. Fossils are char-
acteristically £attened and preserved as impres-
sions in the carbonate matrix or coated by other
Phosphatisation can result in the replication of
soft-tissues over a period of weeks or months.
Apatite forms more rapidly and yields higher ¢-
delity of detail at the cellular level than any other
mineral (Duncan and Briggs, 1996). Phosphatisa-
tion is more prevalent in organisms such as ar-
thropods, ¢shes and cephalopods (Wilby and
Briggs, 1997) that have a relatively high reserve
of organically bound phosphate in the carcass
(Fig. 6D,E). This process frequently occurs in ma-
rine environments, particularly in plattenkalks, as
in the Upper Jurassic limestones of Solnhofen,
Germany, and Cerin, France, the Upper Creta-
ceous Hakel and Hjoula limestones of Lebanon,
and Eocene limestones of Monte Bolca, Italy
(Wilby and Briggs, 1997). However, such deposits
do not always yield insects, although the Solnho-
fen Limestone is an exception (Frickhinger, 1994 ;
Viohl, 1998). In contrast, a number of important
freshwater plattenkalks that do yield insects, in-
cluding the Lower Cretaceous of Montsec and
Las Hoyas, Spain, and the Eocene Green River
Formation of the Western Interior of the USA,
do not normally preserve phosphatised soft-tis-
Phosphatisation of insect soft-tissues normally
requires a build-up of phosphorus in the upper
layers of the sediment in order to provide more
than is available within the carcass itself. This
may be accomplished by adsorption to clays and
iron hydroxides (Wilby et al., 1995). Phosphorus
may be derived from decay of organic matter,
dissolution of biogenic apatite in skeletons, and
from the water column (Wilby et al., 1996). In
order to be available for incorporation into fossils
phosphorus must be prevented from recycling
back into the water column. Microbial mats
may achieve this by controlling the redox bound-
ary, often as a site of transitory ¢xation (Soudry,
1992; Reimers et al., 1988). Microbial mats pre-
vent carcasses from £oating and accordingly pro-
tect them from scavengers and currents (Gall et
al., 1994). These mats also favour early minerali-
sation because they create low oxygen conditions
and reduced pH around the carcass (Chafetz and
Buczynski, 1992; Neumeier, 1999). Mineralisation
of soft-tissues may take several weeks, or even
months, if decay is inhibited (Briggs et al.,
1993). Microbial mats are evident in association
with fossil insects at several localities (see Section
Pyrite is ubiquitous in organic-rich, ¢ne-grained
sediments. It forms in both marine and freshwater
anoxic settings (Allison, 1988), but is usually sul-
phate-limited in the latter due to low sulphate
concentration in freshwater. It is rare in continen-
tal limestones, where most fossil insects occur.
Framboids and crystal aggregates are the most
common textures observed in pyritised insects
(Pen alver et al., 1993). Insects in the Lower Creta-
ceous Crato Formation of Brazil are preserved in
two and three dimensions in goethite (iron oxide
hydroxide; Fig. 7C)(Grimaldi and Maisey, 1990).
The exceptional detail (Grimaldi and Maisey,
1990; Martill, 1993) re£ects mineralisation prior
to signi¢cant decay. These sediments were depos-
ited in a lake that evolved under arid or semiarid
tropical conditions, with a possible seasonal mon-
soon climatic regime that may have favoured per-
manent strati¢cation of the water body. Rapid
burial may have been due to major sediment input
during particular events, or by high sedimentation
rates of s1 cm/yr associated with a delta (Lopes-
Neumann, 1999). A low organic matter content,
evidenced by a matrix that is 99% calcite (Lopes-
Neumann, 1999), together with the presence of
pore-water sulphates, favoured pyrite precipita-
tion during insect decay and decomposition. San-
tana insects may have been mineralised initially in
pyrite and then altered to goethite during diagen-
(3) As a mould. In preservation as moulds only
an impression of the external characters of the
insect remains (Fig. 7F). Where the tissues of
the insect decay rapidly, mould formation re-
quires early diagenetic cementation of the sur-
rounding lithology. More decay resistant cuticles
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^64 45
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^6446
may survive lithi¢cation of the sediment, only to
be removed during diagenesis or weathering. The
result may be an external mould. External moulds
of insects commonly occur in limestones with a
low organic matter content (Mart|
al., 1995), in travertines (Nel and Blot, 1990), and
in carbonate nodules (Barro
¤n et al., 2002).
(4) In travertine. Travertines are freshwater car-
bonates that form in £uvial settings or springs
with a high alkalinity. The major sources of car-
bonate in such settings are inorganically precipi-
tated carbonate, carbonate precipitated as a result
of photosynthesis, biogenic debris from calcare-
ous plants and animals, and allochthonous mate-
rial from carbonate rocks in the drainage basin
(Dean and Fouch, 1991). Precipitation of CaCO3
on the surface of organic remains results in their
preservation as moulds. While plants are often
preserved in such settings, insects are rare, excep-
tions being a dragon£y wing from the Holocene
of Auriol, France (Fig. 7G;Papazian and Nel,
1989), and insects in the Miocene of Bo
Germany (Zeuner, 1931). Traces of insect activity
are even more rarely preserved in travertines, such
as galls produced by the cecidomyid £y Harman-
diola in leaves of Populus, from the Pleistocene of
the Matarranya River in Teruel, Spain (Pen alver
et al., 2002b), and the feeding in wood by a Qua-
ternary beetle of the family Scolytidae, from
«ringens, Germany (Claus, 1958).
(5) In nodules. Insects are not often preserved in
concretions. Examples include the famous Upper
Carboniferous biotas of Mazon Creek, Illinois,
USA, and Montceau-les-Mines, France, and the
Middle Miocene biotas of the Barstow Formation
of the Mojave Desert, California, USA, and of
Izarra, A
Łlava, Spain (Figs. 6I and 7H). The Bar-
stow Formation insects and other arthropods are
preserved in three dimensions and are typically
replaced by minerals, such as quartz (Fig. 7B),
apatite, celestine, gypsum and zeolites, but in
some cases the structure of the soft-tissues of
these organisms is preserved (Palmer, 1957 ;
Park, 1995; Park and Downing, 2001). The in-
sects of Montceau-les-Mines (France) and Izarra
(Spain) are preserved as moulds (Caurel et al.,
1994; Barro
¤n et al., 2002). 3-D preservation indi-
cates that nodules are formed by early diagenetic
processes, prior to decay collapse. In these local-
ities insects do not always occur in the centre of
the nodules. Nodule formation begins with decay
and release of CO2and NH3. This promotes an
increase in concentration of the bicarbonate ion
3) promoting CaCO3precipitation around
the carcass (Berner, 1968; Raiswell, 1987; Rais-
well and Fisher, 2000). Calcareous nodules from
the Montceau-les-Mines basin are dominated by
siderite and pyrite with traces of phosphate; they
may be traversed by calcite veins. This suggests an
anoxic setting rich in organic matter, with a pH of
7 or lower (Caurel et al., 1994). The Miocene
Barstow Formation concretions, on the other
hand, were formed under dysaerobic, low energy
conditions in a distal saline^alkaline lake (Park,
1995). No geochemical studies have been carried
out in Izarra.
Colour is rarely preserved in insect fossils be-
cause the majority of chemical compounds that
cause it are lost rapidly after the death of the
organism. The oldest preserved colour in insects
is known from several Upper Carboniferous de-
posits associated with coals (Carpenter, 1992). Ex-
Fig. 8. (A) Hymenoptera with fungal covering. EPGM-RD-0055; Miocene amber, Hispaniola ; scale bar = 1 mm. (B) Brachycer-
ous Diptera showing a white aureole produced by decay and decomposition of labile tissues. This aureole is best developed on
the side of the amber where the insect was trapped initially. EPGM-B-0056; Eocene, Baltic Sea (Lithuania); scale bar = 0.5 mm.
(C) Spermatophore (black arrow) at the apex of the abdomen of a tiny wasp (Mymarommatidae, Hymenoptera). MCNA 9127;
Lower Cretaceous amber, Moraza (Spain); scale bar = 0.05 mm. (D) Detail of copal stalactite showing cracking of the external
surface that may a¡ect insect preservation. EPGM-CM-0054; Holocene, Madagascar; scale bar = 0.5 mm. (E) Rugose and darker
crust (black arrow) produced by weathering of amber; crystalline amber (white arrow), EPGM; Lower Cretaceous, Escucha
(Spain); scale bar = 5 mm. (F) Wonnacottella pulcherrima, palaeontinid (Homoptera) showing colour pattern in wings preserved
in carbonate. LP-0026-G/IEI; Barremian, El Montsec (Spain) ; scale bar = 10 mm. (G) Xestocephalus, leafhopper showing colour
pattern. EPGM-RD-0057; Miocene amber, Hispaniola; scale bar=10 mm.
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^64 47
amples of colour preserved on insects are known
mainly from carbonate rocks (Fig. 8F) but also
from amber (Fig. 8G). In some Mesozoic dragon-
£ies from carbonate rocks the wing venation pat-
tern is only preserved where the wings were col-
oured; wing zones that lacked colour are not
preserved. This suggests that the properties of
the coloured wings in dragon£ies, perhaps the
presence of melanin, may have promoted preser-
vation. In other groups of insects that preserve
colour, such as leafhoppers and scorpion£ies,
the venation is preserved throughout. The metallic
colours which are preserved in some beetles from
the Eocene of Messel, Germany, for example, are
not chemical in origin, but a product of multilayer
re£ectors (Parker and McKenzie, 2003).
6.3.2. Diagenesis in resins
Processes such as dehydration and carbonisa-
tion begin as soon as the insect is trapped but
before the resin is buried. It is not always clear
which processes occur before and after burial of
the resin so here we consider the diagenesis of
insect inclusions to begin when they are ¢rst
trapped, whereas diagenesis of the resin only be-
gins when it is buried in a sedimentary deposit
(Fig. 6). Resins have antiseptic and antibiotic
properties that protect an insect carcass from de-
cay, unlike those in limestones. The development
of fungi around carcasses in amber where the hy-
phae clearly are not oriented by the £ow of the
resin indicates that the fungus grew after the in-
sect was trapped; thus in some cases the resin
does not have fungicidal proprieties and may
not harden quickly (Fig. 8A). Resin-encapsulated
insects provide an opportunity to investigate dia-
genetic changes in organic matter that has been
protected from external agents.
Insect remains in copal and amber are usually
restricted to the cuticle. After the insect is em-
bedded in resin, anaerobic degradation occurs as
a result of autolysis and the activity of endoge-
nous bacteria (Allison and Briggs, 1991a,b). These
processes may be inhibited by rapid dehydration,
however, and labile tissues are sometimes pre-
served. Experiments have shown that early dehy-
dration is critical to this process (Henwood,
1992a,b;Grimaldi et al., 1994). Henwood
(1992b) argued that insects with well preserved
internal organs represent individuals that dehy-
drated prior to complete entombment. Poinar
and Hess (1982) suggested that such mummi¢ca-
tion might be the result of an extreme case of inert
dehydration. They considered that organic com-
pounds such as sugars and terpenes in the resin
might combine with water in the tissues to aid in
the dehydration process. Grimaldi et al. (1994)
argued that the role of terpenes is probably
more signi¢cant than that of monosaccharides
and other compounds, because terpenes occur in
higher concentrations and would perfuse tissues
more rapidly. During early diagenesis, £uids pro-
duced by decay and decomposition of labile tis-
sues may react with sugars and terpenes in the
resin, resulting in a white aureole around the car-
cass (Fig. 8B) that has been interpreted as a foam
of microscopic bubbles (Mierzejewski, 1978 ;
Weitschat and Wichard, 1998). This frequently
occurs in insect larvae with a high proportion of
labile tissues, but also in adults, particularly in
Baltic and Dominican amber. Where foam is
present on only one side of a carcass the other
probably was exposed to the atmosphere long
enough to become dehydrated before it was cov-
ered by a subsequent £ow. Grimaldi et al. (2000)
reported New Jersey amber that is turbid due to
foams. If soft part preservation is a function of
dehydration, reduced evaporation underground
might contribute to the lower preservation rate
of subterranean organisms, but this is more likely
a re£ection of their low mobility compared to
£ying insects. Internal soft-tissues are preserved
in both Dominican and Baltic amber (Grimaldi
et al., 1994), but more commonly in the former,
which sometimes yields cellular details, including
mitochondria and nuclei (Poinar and Hess, 1982;
Henwood, 1992a). Membranous structures, such
as musculature, air sacs or tracheae, are generally
better preserved than proteinaceous ones (Grimal-
di et al., 1994). Internal soft-tissues also have been
reported from Lebanese amber (Azar, pers. com-
mun.). Gas exchange with the atmosphere may
occur (Hopfenberg et al., 1988; Cerling, 1989),
but this is unlikely to promote decay.
Some macromolecules are more decay-resistant
than others. DNA is highly susceptible to both
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^6448
hydrolysis and oxidation (Lindahl, 1993) and is
rare in fossils more than 100 000 years old. Re-
ports of DNA from insects in amber (Desalle et
al., 1992; Cano et al., 1993) probably re£ect con-
tamination (Austin et al., 1997; Walden and
Robertson, 1997; Cooper and Poinar, 2000).
Proteins are known to decay rapidly ; some of
them, such as albumin, have similar characteris-
tics to DNA (Poinar et al., 1996), while others,
such as collagen, osteocalcin or keratin, are more
resistant (but see Bada et al., 1999). Cellulose and
chitin are carbohydrates that are easily biode-
graded unless they are part of structural tissues
through cross-linking with other molecules. Rec-
ognisable traces of these biomolecules may last
for millions of years (Briggs et al., 1998 ; Briggs,
Low molecular weight fractions such as mono-
and sesqui-terpenes readily di¡use from the resin
into carcasses. Insect preservation in Baltic (gym-
nospermous origin) and Dominican (angiosper-
mous origin) ambers di¡ers (Grimaldi et al.,
1994). Dominican amber usually preserves mem-
brane structures, musculature, and nerve tissue,
whereas preservation of these features is less fre-
quent in Baltic amber where inclusions tend to be
surrounded by a milky coating due to microbial
decomposition, autolysis or reaction between de-
cay £uids and compounds in the resin. Baltic am-
ber contains a high concentration of resin acids,
mainly succinic and communic. Dominican and
Mexican ambers are similar to Baltic amber and
fossil kauri, but di¡er in the stereochemistry of
the methyl group and its side-chain. Thus they
are not derived from communic acid but from
ozic acid and its hydroxy derivative zanzibaric
acid (Beck, 1999). The preservation of amino
acids in insect tissues from ambers of various
ages is attributed to anhydrous conditions
(Wang et al., 1995). Low temperatures are a
more important factor in determining whether
amino-acids might be preserved in amber than is
the age of the deposit (Halpine, 1995; Wang et
al., 1995). Traces of chitin and protein are present
in bees from 2000-year old African copal, but
bees and beetles from Dominican amber preserve
neither. The cuticle of the Dominican taxa has
been transformed into long, straight chain hydro-
carbons (Stankiewicz et al., 1998b). Encapsulation
of the insects in resin precludes an external source
for the aliphatic polymer, which is interpreted as a
diagenetic product of polymerisation of the insect
cuticular waxes and body lipids (Stankiewicz et
al., 2000).
Resin prevents carcasses from collapsing and
allows them to retain their 3-D morphology.
The preservation of decay-prone tissues is doubt-
less a product of reaction with constituents in the
resin, although the details of this process remain
to be elucidated. Henwood (1993) suggested that
diterpene resin acids, such as copalic and enantio-
pinifolic acids, reacted with organic inclusions.
Stankiewicz et al. (1998b) considered the preser-
vation to be a result of cross-linkage with resin
components connected via sulphur and/or other
terminal ole¢nic groups of the diterpenoid units.
The loss of volatiles during polymerisation of co-
pal under atmospheric conditions promotes crack-
ing of the surface (Fig. 8D).
Amber may be reworked. Amber deposits are
known to have been eroded and redeposited in
the Baltic Sea region during the Quaternary (Ma-
sicka, 1972; Weitschat and Wichard, 1998). Baltic
amber even has been found in the UK (Jarzem-
bowski, 1999). The ease with which resin pieces
are transported varies with their density, which
depends on the nature of the inclusions and de-
gree of polymerisation, and with the salinity of
the transporting medium. The density of resin is
similar to that of water (copal: 1.03^1.08 ; amber :
1.04^1.10). Dominican amber shows little evi-
dence of reworking or transport and its age can
be used to calibrate rates of molecular evolution
of the included taxa (Iturralde-Vinent and Mac-
Phee, 1996).
Amber oxidises rapidly when exhumed, indicat-
ing, together with the results of studies of Recent
kauri gum (Poinar, 1991a), that preservation is
favoured by burial in anoxic sediment. The pro-
cesses that a¡ect amber-bearing rocks also a¡ect
their inclusions. Darkening of amber, usually
from yellow to red, has been correlated with ox-
idation and weathering, as well as temperature
increase: amber from both the Baltic (Eocene)
and northern Spain (Middle Cretaceous) exhibits
a darkening of the edges of abundant organic in-
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^64 49
clusions. A comparison of the maturity of amber
with that of the enclosing rock matrix, as deter-
mined by the study of the vitrinite content, may
reveal whether or not the amber has been re-
worked (Poinar and Mastalerz, 2000).
The major processes that a¡ect amber-bearing
deposits during diagenesis are overburden pres-
sure and elevated temperature. Intensely polymer-
ised amber is brittle, while less polymerised amber
is more plastic and deformable and may become
£attened or disc-shaped (Zherikhin and Eskov,
1999). Release of pressure may result in fracturing
around the discontinuity formed by an insect in-
clusion. Fractures and haloes around carcasses
are common, particularly in Baltic amber. Dark
amber layers, which have been polymerised by
sunlight and wind, are more prone to fracturing.
Increased temperature may induce darkening.
Moreover, amber may melt (melting point : 200^
380‡C) or fragment completely. Di¡erences in de-
gree of tectonic deformation are re£ected in di¡er-
ences in the infrared spectra of amber from two
localities in the Cordillera Septentrional of the
Dominican Republic: La Toca and Palo Alto
(Henwood, 1992a). The La Toca site has under-
gone tectonic deformation (Grimaldi, 1995).
Weathering destroys amber through oxidation,
which results in the formation of a friable crust
that increases in thickness and darkness with time
(Fig. 8E). Diagenetic minerals, such as pyrite, oc-
cur in some ambers. Pyrite may penetrate amber
to form dendrites along ¢ssures, or precipitates as
a coat at the boundary between the amber matrix
and the insect carcass (Schlu
«ter and Stu
1982; Baroni-Urbani and Graeser, 1987). Pyrite
most commonly occurs within amber in sulphur-
rich coal-bearing sediments (Garty et al., 1982 ;
Alonso et al., 2000). Few examples are known
of completely pyritised insects: ¢ve specimens in
Lebanese amber (Azar, pers. commun.), and two
in New Jersey amber (Grimaldi et al., 2000). The
mode of formation is unknown, but is likely to
re£ect microfractures that connect the insect to
the surface of the amber piece. Schlu
«ter (1989)
found microscopic crystals of marcasite in the in-
terior of some insects from Cenomanian amber of
northwestern France.
Collecting methods may introduce biases into
insect assemblages and distort palaeoecological
conclusions. Pike (1993) compared the insects
from ¢ve samples of amber that were prepared
by picking with those from three samples that
were prepared by sieving. Fewer insects per kg
were found in the picked samples than in the
sieved ones. Sieving evidently retrieved small
pieces containing signi¢cant inclusions. Amber
preserves many small insects such as thysanopter-
ans and aphids, but this size range is very rare in
limestones. This bias may be partly a result of
collecting methods, as limestones are usually
sampled in the ¢eld, whereas amber is sampled
using a microscope in the laboratory.
7. Conclusions
Localities that yield fossil insects are uncom-
mon compared, for example, to those preserving
mollusc shells: insects are treated as an exotic
group of fossil arthropods in the majority of pa-
laeontology texts, and are sometimes not men-
tioned. In recent years, however, a large number
of papers have been published on fossil insects:
they are the fossil group to have received the most
attention between the publication of The Fossil
Record in 1967 (Crowson et al., 1967) and The
Fossil Record 2 in 1993 (Ross and Jarzembowski,
1993; Labandeira, 1994). Today more species of
fossil insect are known than any other group of
fossil arthropods, and there are more families of
fossil insects than of combined fossil and recent
The fossil record of insects preserved in carbon-
ate rocks is very di¡erent from that in amber.
Evidence of the death of the insect is usually pre-
served in amber, but not in carbonate rocks. Bio-
stratinomic processes are much more signi¢cant in
sediments, although insects in amber may be
transported together with the preservation me-
dium. The diagenesis of insects entombed in resin
is poorly understood, but dehydration, reactions
with chemical components within the resin, and
polymerisation are important. The diagenesis of
insects found in carbonate rocks is more complex
and diverse. Specimens may be preserved as or-
ganic remains, albeit altered to a more stable
composition; by replication or replacement in an
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^6450
authigenic mineral, usually calcium carbonate, py-
rite or calcium phosphate ; as an external mould
of the exoskeleton; or as a cast in calcite, silica, or
pyrolusite. Insects are usually £attened in lime-
stones, while they are preserved in three dimen-
sions in amber.
Although fossil insect localities are rare and
scattered in time and space, where these arthro-
pods do occur, they tend to be abundant and
diverse. Few insect localities are known from the
Triassic, Upper Cretaceous and Palaeocene. Thus
the insect fossil record £uctuates with global ta-
phonomic conditions. Family level data across the
Cretaceous^Tertiary (K^T) boundary reveal little
impact of the extinction on insects (Whalley,
1987; Ross et al., 2000; but see Christianson
and Pike, 2001), but no major occurrences of in-
sects of Maastrichtian or Danian age are known
(Labandeira et al., 2002). Evidence of plant^insect
associations, however, suggests that there was a
major extinction of insect herbivores at the K^T
boundary, at least in North America (Labandeira
et al., 2002). Fluctuations in the quality of the
insect fossil record generate problems for the
study of fossil insects: in analyzing phylogeny
(some groups are less represented than others) ;
in palaeobiogeography (e.g. the lack of insect re-
cords for most of Gondwanaland); in determining
the true timing of the origin and extinction of
some groups; and in identifying the history of
diversity of the group as a whole and the timing
of important radiations and extinctions, such as
the impact of the Permo^Triassic and K^T
boundary events.
This review of the major taphonomic processes
that a¡ect the preservation of insects in carbon-
ates and amber will assist in the identi¢cation of
the factors that promote biases in the insect fossil
record. It also highlights questions that require
further research. Why are biological inclusions
absent at most amber localities and how was
this barren amber produced ? How is amber trans-
ported from the producing tree to the sediment in
which it is preserved? What controls 3-D preser-
vation and concretion formation in carbonates ?
The study of taphonomic processes is impor-
tant not only for the reconstruction of the fossil
insect assemblage, but also for the interpretation
of the sedimentary setting and palaeoenvironment
where the insect lived (climate) and died (anoxia,
presence of microbial mats), as well as the ecolog-
ical importance of insects in Phanerozoic terres-
trial ecosystems. Reworking of amber may result
in time-averaging and mixing of insects from
di¡erent ecosystems. An understanding of the
taphonomy of insect assemblages is essential to
determining the palaeoenvironment and interpret-
ing possible relationships among insects, interac-
tions between insects and other organisms, and
their role in ancient ecosystems.
We thank D. Azar, C. Mart|
¤n-Closas, and A.
Nel for comments on an early version of the
manuscript, V.V. Zherikhin for information about
Russian localities, and S. Elias and particularly
C.C. Labandeira for very constructive reviews of
the submitted paper. D.B. Rowley kindly pro-
vided the programme for plotting the palaeogeo-
graphic data. We are grateful for advice and as-
sistance in plotting the maps from A.M. Ziegler
and T.R. Rothfus. Aspects of this research were
supported by the ‘Fossil Insects’ project of the
European Science Foundation and by grants
MCTE (Spain) 2001-0173 and MCTE 2001-
0185. The study of the A
Łlava amber (Basque
Country) was supported by the Diputacio
¤n Foral
de A
Łlava. The contribution by D.E.G.B. was
completed while he was a visiting professor at
the Department of Geophysical Sciences, Univer-
sity of Chicago, and visiting scientist at the Field
Museum of Natural History, Chicago, USA. A.
Arillo (Madrid), M. Belincho
¤n (Vale
'ncia), A. La-
casa (Lleida), A. Nel (Paris), L. Park (Akron), C.
¤n-Closas (Barcelona), M. Solsona (Barcelo-
na), A. Rasnitsyn (Moscow), G. Viohl (Eichsta
J. Wagensberg (Barcelona), and V. Zherikhin
(Moscow), and their respective institutions kindly
provided photographs or access to some of the
specimens that were studied and ¢gured. This pa-
per is dedicated to the memory of Dr. V.V. Zher-
ikhin who suggested that we should write it, and
provided encouragement and advice during its
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^64 51
Amber deposits Age Coordinates Palaeolatitude Palaeolongitude
1 Aukland Prov. (New Zealand) Miocene^Pliocene 37S 175E 344.31 175.411
2 Luzon Island (Phillipines) Miocene^Pliocene 16N 121E 16 121
3 Romania Upper Miocene 44N 26E 43.851 24.761
4 Yazov (Ukraine) Upper Miocene 48N 39E 48.294 37.477
5 Okanagan Highlands (British Columbia) Lower Miocene^Middle Miocene 50N 119W 50.106 3112.591
6 Bayaguana^El Valle (Dominican amber) Lower Miocene 19N 69W 17.512 367.53
7 Merit^Pila (Malaysia) Lower Miocene 2N 113E 4.428 117.304
8 Burghammer (Lausitz, Germany) Miocene 51N 14E 50.437 12.342
9 Central Sumatra Miocene 0 102E 1.157 106.602
10 Macho
¤w (Poland) Miocene 51N 22E 50.705 20.258
11 Para
¤(Brazil) Miocene 5S 53W 36.745 348.228
12 Goitsche mine (Germany) Upper Oligocene^Lower Miocene 52N 12E 51.193 7.935
13 Bitterfeld (Germany) Upper Oligocene^Lower Miocene 52N 12E 51.193 7.935
14 Catania (Sicily) Upper Oligocene 37N 15E 34.338 14.153
15 Santiago^P. Plata (Dominican amber) Middle Oligocene^Miocene 19N 71W 15.54 368.83
16 Mexican amber (Chiapas^Simojovel) Middle Oligocene^Miocene 17N 93W 14.081 383.643
17 Chilikty Fm. (Pavlodar, Kazakhstan) Middle Oligocene 52N 77E 56.044 75.29
18 Chulym^Enisey basin (Siberia) Middle Oligocene 58N 85E 62.376 83.34
19 Rechnoy Peninsula (Tavrichanka, Russia) Oligocene 43N 132E 47.39 134.406
20 Uly^Zilansik River Basin (Kazakhstan) Oligocene 49N 65E 52.385 62.729
21 Iwaidzumi^Iwate (Japan) Oligocene 40N 142E 48.971 143.454
22 Korzhindy Fm. (Aral Sea, Kazakhstan) Lower^Middle Oligocene 45N 60E 48.08 58.016
23 Kantemirovka (Veronezh, Russia) Lower Oligocene 50N 40E 51.549 36.338
24 Colti (Buzau, Romania) Early Oligocene 45N 27E 45.507 24.102
25 Ben Metir (A|
«n Draham, Tunise) Oligocene or earlier 37N 9E 34.227 8.356
26 Klesov (Ukraine) Upper Eocene^Lower Oligocene 51N 27E 51.356 21.207
27 Uglovka Fm. (Primorye, Russia) Eocene^Oligocene 55N 20E 54.267 12.89
28 Seattle (Washington, USA) Eocene^Oligocene 48N 122W 45.767 3101.135
29 Avekova Fm. (Magadan, Russia) Eocene^Oligocene 60N 151E 65.524 166.927
30 Ko
«nigsaue (Aschersleben, Germany) Upper Eocene 52N 11E 50.119 5.513
31 Palmnicken (Samland, Russia) Upper Eocene 55N 20E 54.267 12.89
32 Zwenkau (Leipzig, Germany) Upper Eocene 51N 12E 49.277 6.786
33 Dnepr River basin (Belorussia and Ukraine) Upper Eocene 47N 34E 48.435 29.504
34 Kasyanovo (Veronezh, Russia) Upper Eocene 50N 40E 52.205 34.602
35 Verkhne^Sinevidnoe (Ukraine) Upper Eocene 49N 20E 48.422 15.069
36 Sambian Peninsula Upper Eocene 55N 20E 54.267 12.89
37 Parczew (Poland) Early Upper Eocene 52N 23E 51.768 16.942
38 Shansi Prov. (China) Middle^Upper Eocene 35N 112E 43.043 114.627
39 Shensi Prov. (China) Middle^Upper Eocene 34N 109E 42.17 111.203
40 Kwangtung (China) Middle^Upper Eocene 23N 113E 30.459 118.924
41 Yungchang (Yunnan, China) Middle^Upper Eocene 26N 100E 30.79 103.126
42 Tiger Mountain Fm. (Issaquah, Washington, USA) Middle Eocene 48N 122W 45.767 3101.135
PALAEO 3225 5-1-04
X. Mart|
's et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 19^6452
Appendix (Continued).
Amber deposits Age Coordinates Palaeolatitude Palaeolongitude
43 Claiborne Fm. (Arkansas, USA) Middle Eocene 36N 92W 31.062 374.884
44 Napan Fm. (Kamchatka, Russian) Middle Eocene 58N 159E 62.63 175.052
45 Visloe (Belgorod, Russia) Middle Eocene 51N 37E 52.763 31.185
46 Helmstedt (Germany) Middle Eocene 52N 11E 50.119 5.513
47 Leonhard (Geiseltal, Germany) Middle Eocene 51N 12E 49.277 6.786
48 Baltic amber (Samland Peninsula) Lower^Middle Eocene 55N 20E 54.267 12.89
49 Guchenzgi Fm. (Fushun, Liaoning, China) Eocene 42N 124E 48.61 137.14
50 Malverne (Arkansas, USA) Eocene 34N 93W 29.148 376.044
51 Simi Valley (California, USA) Eocene 34N 119W 35.089 3102.457
52 Tastakh Lake (Siberia) Eocene 71N 144E 77.457 164.343
53 Patagonica Fm. (Puerto Madryn, Argentina) Lower^Middle Eocene 43S 65W 350.66 341.917
54 Latdorf^Nienburg (Germany) Eocene 52N 12E 50.254 6.451
55 Nietleben (Halle, Germany) Eocene 51N 12E 49.277 6.786
56 Ameki Fm. (Umuahia^Bende, Nigeria) Eocene 6N 7E 1.352 6.924
57 Eureka Sd Grp (Axel Heiberg+Ellesmere Island,
Eocene 80N 85W 73.568 353.205
58 Gablitz (Wien, Austria) Eocene 48N 16E 46.89 11.532
59 Douzens (Aude, France) Lower Eocene 43N 3E 40.239 0.512
60 Quesnoy (Oise, France) Lowermost Eocene 51N 3E 46.543 11.429
61 Hat Creek (British Columbia, Canada) Early^Middle Eocene 51N 121W 48.62 399.467
62 Nanaimo (Vancouver, Canada) Early^Middle Eocene 49N 124W 46.969 3102.794
63 Princeton (British Columbia, Canada) Early^Middle Eocene 49N 121W 46.647 399.957
64 Bykovskaya Fm. (Lena Delta, Siberia) Palaeocene 72N 127E 73.946 101.574
65 Naiba River (near Urtay Creek mouth, Russia) Palaeocene 52N 143E 59.508 148.527
66 Sakhalin Island (Russian Far East) Palaeocene 49N 143E 57.192 144.891
67 London Clay Fm. (UK) Palaeocene 51N 1E 41.722 7.473
68 Wabamun Lake (Alberta, Canada) Palaeocene 54N 114W 61.695 374.752
69 Corbie
'res (Pirenean, France) Danian 43N 2E 33.705 8.015
70 Lance Fm. (Niobrara, Wyoming, USA) Upper Maastrichtian 43N 104W 51.426 359.433
71 Hell Creek Fm. (Glendive, Montana) Upper Maastrichtian 47N 105W 55.375 357.831
72 Laramie Fm. (Boulder, Colorado) Upper Maastrichtian 40N 105W 48.921 362.264
73 Timmerdyankh^Khaya Hill (Yakutia, Russia) Maastrichtian^Danian 66N 174E 74.869 162.467
74 Edmonton (Alberta, Canada) Maastrichtian 54N 113W 63.899 360.676
75 Grande Prairie (Alberta, Canada) Maastrichtian 55N 119W 66.419 366.885
76 Vilada (Barcelona, Spain) Maastrichtian 41N 2E 28.713 13.446
77 Terlingua Formation (Brewster County, Texas) Late Campanian^Early Maastrichtian 30N 103W 38.951 364.788
78 Fruitland Fm. (San Juan Basin, New Mexico) Campanian 37N 107W 46.362 366.119
79 Cedar Lake (Manitoba, Canada) Campanian 53N 100W 59.264 347.468
80 Foremost Fm. (Medicine Hat, Alberta, Canada) Campanian 50N 111W 59.775 362.183
81 Baja California (Mexico) Campanian V29N 115W