Content uploaded by Edyta J. Jasinska
Author content
All content in this area was uploaded by Edyta J. Jasinska on Mar 19, 2016
Content may be subject to copyright.
Root Mats in Ground Water: A Fauna-Rich Cave Habitat
Author(s): Edyta J. Jasinska, Brenton Knott and Arthur J. McComb
Source:
Journal of the North American Benthological Society,
Vol. 15, No. 4 (Dec., 1996), pp.
508-519
Published by: Society for Freshwater Science
Stable URL: http://www.jstor.org/stable/1467802 .
Accessed: 03/10/2013 13:14
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
.
Society for Freshwater Science is collaborating with JSTOR to digitize, preserve and extend access to Journal
of the North American Benthological Society.
http://www.jstor.org
This content downloaded from 142.244.226.141 on Thu, 3 Oct 2013 13:14:40 PM
All use subject to JSTOR Terms and Conditions
J.
N. Am. Benthol.
Soc.,
1996,
15(4):508-519
e 1996
by The North American
Benthological Society
Root mats in ground water: a fauna-rich cave habitat
EDYTA
J. JASINSKA
Environmental
Science,
Murdoch
University,
Murdoch,
Western
Australia 61501
BRENTON KNOTT
Department of Zoology, The University of Western
Australia,
Nedlands, Western
Australia 6907
ARTHUR J. MCCOMB
Environmental
Science,
Murdoch University,
Murdoch,
Western Australia 6150
Abstract. An unusually
rich cave fauna of 41 species
is reported
from a 20-m stretch of a shallow
groundwater
stream in Western
Australia.
Epigean
waters do not occur upstream
of the cave and,
downstream,
cave waters emerge at the surface after flowing about 100 m through a system of
limestone caverns and collapses. Most of the aquatic
cavernicoles
collected from the cave stream,
therefore,
appear to be confined to the subterranean
realm. Potential
food sources for the aquatic
fauna
of the cave
were investigated.
Inputs
of dissolved organic
matter,
surface
debris, cadavers,
and
guano into the cave stream
were lacking
or too small to provide sufficient
energy for the abundant
aquatic
cavernicoles.
Aquatic
root
mats were shown to be the only substantial source
of food available.
They belong to the native tree Eucalyptus gomphocephala
and consist of fine, highly branched,
feeder
rootlets
growing
in mycorrhizal
association
with fungi. Most of the animals,
both in terms of density
and species
number,
also occur
in the submerged
root mats. At least some of the aquatic
cavernicoles
graze directly
on the living rootlets,
some feed on root mat detritus,
and others are predatory.
We
propose that tree roots provide an important
source of energy for subterranean fauna in shallow
groundwaters
both in Australia
and elsewhere.
Key
words: biospeleology,
cave habitat,
subterranean
water, stygal habitat,
groundwater
fauna,
stream, roots, aquatic
root mats.
The dark zones of cave environments typical-
ly contain few species and low numbers of an-
imals when compared with surface (epigean)
ecosystems in the same area (e.g., Barr 1967,
Culver 1982). The main reason usually ad-
vanced for this impoverished state of cave fau-
nas is lack of a nutritious, substantial and reli-
able food source; this lack is, in turn, attributed
to the absence of photosynthetic primary pro-
duction, and the isolation of caverns from sur-
face environments. The richest cave ecosystems
reported to date have substantial inputs of or-
ganic energy sources such as combinations of
bat and cricket guano, cadavers, washed-in or-
ganic matter, and tree roots (e.g., Barr 1967,
Longley 1981, Culver 1982, Howarth 1983), one
notable exception being a chemoautotrophically
driven aquatic cave fauna of Movile Cave in Ro-
mania (Sarbu et al. 1996). Tree roots in caves
represent the only known instance of live plants
1
Present address:
Department
of Zoology,
The Uni-
versity of Western
Australia, Nedlands,
Western Aus-
tralia 6907.
providing a potential food source for caverni-
coles in the dark zones of caves. Aerial tree roots
have been recognised
as an important
source of
energy for terrestrial
cave insects in a number
of tropical
caves in Queensland
(Howarth
1988)
and Hawaii (Howarth 1972).
However,
aquatic
root habitats
in caves have received little atten-
tion, despite published reports
of collections of
cavernicoles from submerged
root mats in caves
(e.g., Notenboom
1981).
Here
we report
on an aquatic
root mat habitat
and associated fauna in Cabaret
cave, a small
limestone cave in Yanchep
National
Park,
south-
western Australia
(Fig. 1). Before
1990, 12 spe-
cies of animals (Osteichthyes: 1 sp.; Insecta 2
spp.; Crustacea: 6 spp.; Turbellaria: 1 sp.; Hir-
udinea 1 sp.; Oligochaeta:
1 sp.) were reported
from the stream in Cabaret
cave, then referred
to as Silver Stocking
(Burt 1982).
We present a
study of Cabaret cave stream environment, and
attention is particularly given to the food
sources available to the aquatic cavernicoles. The
study was conducted in 1990-1991 and is de-
508
This content downloaded from 142.244.226.141 on Thu, 3 Oct 2013 13:14:40 PM
All use subject to JSTOR Terms and Conditions
ROOT-MAT HABITATS IN CAVE
WATERS
FIG.
1. (A) Yanchep
National
Park
showing the location of Cabaret
cave and other
caves with aquatic
root
mats. (B)
Location of Yanchep
National Park in relation
to water table
contours
of the Gnangara
Mound
aquifer.
(C) Location of the study area.
scribed in detail in Jasinska (1990) and Jasinska
and Knott (1991).
Study Site
General
area
The climate of the study area is Mediterra-
nean with cool wet winters and hot dry sum-
mers. Yanchep National Park lies between
31?30'S and 31?35'S, and 115?39'E and 115?43'E
(Fig. 1). The Park covers a large area of aeolian
calcarenite which is typically 3-20 m thick and
rests on quartz sands of early to mid-Pleistocene
age (Kendrick et al. 1991). The quartz sands con-
tain the "Gnangara Mound" (Fig. 1), a shallow,
unconfined aquifer extending from about
31015'S to 32?00'S, bounded by the Indian
Ocean to the west and by the Darling Scarp to
the east. Where the water table coincides with
the boundary of sands and overlying calcareni-
te, extensive cave systems occur and, in several
caves, subterranean streams have formed at the
water table surface (epiphreatic streams). Five
caves with such permanent epiphreatic streams
in Yanchep National Park contain fauna-rich
root-mat habitats as described below. The root
mats are produced.by native tuart trees (Euca-
lyptus gomphocephala)
which grow above the
caves. The most extensive aquatic root mats oc-
cur in Cabaret cave.
Cabaret cave
Cabaret cave is shallow and small, the cave
ceiling being separated from the epigean
ground surface by less than 1 m of soil (mainly
1996] 509
This content downloaded from 142.244.226.141 on Thu, 3 Oct 2013 13:14:40 PM
All use subject to JSTOR Terms and Conditions
E. J. JASINSKA
ET AL.
direction of stream flow
exit and entry of stream
from Cabaret cave
sediment trap
5-10 cm high barrier
20-25 cm high barrier
I_- *dry cave floor and limestone rocks
n,n root mats
.streambed deposits
bankside pools
2 nl
__
FIG. 2. Cabaret cave stream, showing the distribution of root mats, deposits of fine detritus, sediment traps,
and mesh barriers.
calcarenite, with a veneer of sand and very
poorly developed organic horizon). The epi-
phreatic stream, 8 m below ground, traverses 18
m of Cabaret cave (Fig. 2). The stream is ap-
proximately 2 m wide and 2-3 cm deep with
channels near the stream edges up to 20 cm
deep and 10-20 cm wide. Water level changes
little throughout
the year. When measured on
20 occasions, throughout all seasons, between 15
March 1990 and 3 March 1991, the Cabaret
stream surface fluctuated between 11.15 and
11.17 m Australian Height Datum (AHD) (Jas-
inska and Knott 1991). Flow varies from nearly
stagnant in sheltered mini-embayments to 3
cm/s in shallow mid-stream sections to 21 cm/s
in deeper channels (Jasinska
and Knott 1991).
Patches
of dark
detritus
(hereafter
referred to as
"streambed detritus"), up to 5 cm thick, accu-
mulate in sheltered areas of the stream. Tuart
roots (10-15 cm diameter)
descend
into the cave,
some reaching the stream banks where they
sprout
dense masses of finely branching
rootlets
into the water. These masses of fine roots (0.1-
0.5 mm diameter),
contain much rootlets detri-
tus (confirmed
by light microscopy)
and typi-
cally form "root mats", the upper surfaces of
510 [Volume 15
This content downloaded from 142.244.226.141 on Thu, 3 Oct 2013 13:14:40 PM
All use subject to JSTOR Terms and Conditions
ROOT-MAT HABITATS IN CAVE WATERS
TABLE 1. Comparisons
of some species-rich
of groundwater
faunas.
Data from the
present
study are
included
in the Cabaret cave entry.
Number
of Location/habitat/size
of
species habitat area available Potential food source Reference Comments
6a Samples
from wells inter- Not stated Sket 1991
secting a shallow-fresh-
water
aquifer
in Cyprus
11a Czech (Barrandium)
karst;
bores in shal-
low groundwaters
15-20 Exit Cave
system, Tas-
mania, groundwater
18 Movile
Cave,
southern
Romania;
ground
water and fungal/
bacterial mats
35-40 Edwards
Aquifer,
Texas,
USA, fauna extracted
mainly from wells
Washed
in surface debris Vanek 1982
Surface debris
percolating Eberhard
1992
into the cave;
debris
from aerial
root mats;
faeces and cadavers
from terrestrial
caverni-
coles
Chemoautotrophic
(Hy- Sarbu et al. 1996
drogen sulphide)
bacte-
ria (fungal/bacterial
mats floating
on water
surface)
Energy
source from oil, Longley
1981
peat, washed-in
detritus
mediated into the food
chain via bacteria and
fungi
41 Cabaret
cave,
south-west- Root mats
ern Australia
43 Coastal Plain Aquifer
of Wide
variety
the Yucatan
Peninsula,
Mexico
No comment on sampling
procedure;
2 spp of iso-
pods, 1 sp of amphi-
pod, + Copepoda,
Oli-
gochaeta
and Ostracoda
(not identified
further)
Interstitial fauna of a shal-
low aquifer;
animals ob-
tained by pumping
from
bore holes -10 m
below ground
surface,
bore yields of fauna
ranged from 0 animals
per m3
to 11 spp per m3
(pumped
water)
and up
to 135 animals/m3
(mi-
croinvertebrates)
Over 17 km of passages
and hydrologic
continu-
ity with other
cave sys-
tems in southern Tas-
mania
Species
number
refers to
troglobites
(obligate
cave dwellers),
"sever-
al" species of troglo-
philes (non-obligate
cave dwellers)
also live
in the subterranean wa-
ters. The aquifer
is 282
km long; 8-64 km wide;
122-152 thick
Jasinska
et al., Small
groundwater
present
study stream,
20 m long, less
than 20 cm deep
Reddell 1981 Fauna
obtained
from 157
sites such as caves and
wells
a Groundwater
bore with highest animal abundance
which coincide with that of the stream. The root
mats are not a prominent feature in Cabaret
stream; they are approximately 10 cm thick and
extend only 10-20 cm into the stream channel.
Given the size of the groundwater stream, the
Cabaret fauna is unusually rich compared with
other subterranean waters (Table 1). Such a rich
cave fauna in a small groundwater stream must
1996] 511
This content downloaded from 142.244.226.141 on Thu, 3 Oct 2013 13:14:40 PM
All use subject to JSTOR Terms and Conditions
E. J. JASINSKA
ET AL.
be supported by a substantial and reliable food
source. The potential sources of organic material
for food webs in Cabaret cave stream were as-
sessed in the following ways:
1. Determination of (i) dissolved organics: sug-
ars, amino acids, and organic acids; (ii) debris
from the surface (e.g., leaves, sticks and partic-
ulates carried in rainwater percolating from the
ground surface through the cave ceiling), faeces,
and cadavers; (iii) dead and living tree roots;
and (iv) fungi;
2. Comparison of the abundance and species
richness of fauna in different microhabitats of
Cabaret stream (since high numbers and species
richness of fauna are likely to be indicative of
high food availability in the microhabitat);
3. Comparison of the degree of colonisation by
aquatic cavernicoles of root mats and synthetic
mats of similar dimensions made from fibres
split from polyethylene tape;
4. Direct observation of rootlet grazing by cav-
ernicoles.
Methods
Cabaret water was sampled for physicochem-
ical properties on several occasions between
March 1990 and March 1991, throughout all sea-
sons, and included sampling on 2 consecutive
days. The variables measured were temperature,
pH, total alkalinity, dissolved oxygen, and con-
ductivity.
Investigation
of potential food sources
The only entrance to Cabaret cave is at the
base of a lateral extension of another, partially
collapsed cave, and there are no openings in the
Cabaret ceiling through which animals might
fall. However, during heavy rains, rainwater car-
rying particulate and dissolved organic matter
from the surface may percolate into the cave
through the ceiling. The stream is partially pro-
duced within Cabaret cave by groundwater
emerging from the sandy bed, and partially by
epiphreatic waters which flow into the cave,
through a rock pile, from an upstream system
of caves. It is possible, therefore, that upstream
of Cabaret cave there are areas of cave collapse
where the stream may receive inputs of epigean
matter. The contribution of epigean matter to
the Cabaret cave stream was assessed by install-
ing 1) 15 flexible, plastic, l-mm2-mesh barriers
(to retain large debris) with the lower edge of
each barrier buried in the streambed and the
upper edge at least 2.5 cm above the water level,
and 2) 16 sediment traps (to collect small par-
ticulates) in the stream (Fig. 2). Ten of the 15
barriers, 20-25 cm high and 30-40 cm long,
were set in the deepest water, and 5 barriers, 5-
10 cm high and 30-40 cm long, were installed
in the shallow parts of the stream on 13 March
1990 (Fig. 2). The barriers, together with
trapped debris, were retrieved and immediately
replaced with clean barriers on 17 April 1990
and 20 August 1990; the barriers and retained
debris were collected for the last time on 30 Au-
gust 1990. The 16 sediment traps comprised cy-
lindrical (4 cm diameter, 10 cm deep) vials
placed in the stream bed with the open end at
least 1.5 cm below the water surface, and the
rim of each vial protruding approximately 0.5
cm above the stream bed. Eight sediment traps
were installed under root mats and 8 in the
open water (Fig. 2) on 15 February 1990; their
contents were collected on 30 June 1990 and on
10 August 1990 and examined microscopically
(using 160 to 400 magnifications). Both the
mesh barriers and sediment traps were exam-
ined, but not cleaned, on a monthly basis to
check for macroscopic cadavers and to ensure
that excessive amounts of debris did not become
trapped on or in them. Streambed deposits of
dark, fine detritus and organic matter separated
from cores of the top 4 cm of the stream bed
were subsampled and identified using a com-
pound light microscope (400x magnification).
Levels of common dissolved amino acids,
sugars, and organic acids in Cabaret cave
stream were measured by High Pressure Liquid
Chromatography (HPLC). Measurements for
dissolved organic compounds were undertaken
just once during the 1990 study period, on 6
April 1990. Water samples, 100 mL, were col-
lected from the open stream and root mat in-
terstices, filtered in the field using GF/C (What-
man) filterpaper, concentrated 100X using a ro-
tary evaporator at 30?C, and analysed for amino
acids using Varian 5560 HPLC system and for
sugars and organic acids using BIORAD Cus-
tom Organic Acid and Sugar HPLC System.
To estimate the contribution of new growth of
root mats to the potential food sources for Cab-
aret stream fauna, discrete root-mat branches
were enclosed in cages over 20 wk starting 23
May 1990. The cages, made from plastic mesh
512 [Volume 15
This content downloaded from 142.244.226.141 on Thu, 3 Oct 2013 13:14:40 PM
All use subject to JSTOR Terms and Conditions
ROOT-MAT HABITATS IN CAVE WATERS
4 mm X 5 mm, excluded larger animals capable
of tearing the root mats, e.g., the cave crayfish.
Growth was measured by change in root-mat
area-by photographing each root mat against
a 1 x 1-cm grid at the beginning and at the end
of the experiment.
Fungi associated with root mats would rep-
resent an additional potential food source in the
cave stream. Therefore whole mounts and his-
tological sections of the tuart rootlets were ex-
amined under a compound light microscope
(400x magnification) for the presence of fungi,
using 2 techniques: 1) whole mount squashes of
fresh roots cleared in 10%
potassium hydroxide
and stained with lactic-glycerol-blue (fungal
stain), and 2) fine sections of resin-embedded
rootlets (stained for fungi with Pianese IIIB).
Distribution of
fauna in stream microhabitats
Since high numbers of invertebrates in cave
habitats may indicate high availability of food,
the faunas of 5 Cabaret stream habitats were
compared on 10 August 1990. The sampling
was carried out on 1 date only because more
extensive sampling was likely to have an ad-
verse effect on the animal populations inhabit-
ing not only a small cave stream but also very
small areas within it, especially the bankside
pools (Fig. 2). The 5 habitats were 1) sandy bed;
2) streambed detritus (deposits of dark root de-
tritus in the quieter sections of the stream); 3)
open water; 4) bankside pools on the stream
bank, just above the stream level, and fed from
slow groundwater seepages (these pools are <1
cm deep and 10-20 cm broad, with a loose bed
of dark root detritus and containing a few
strands of rootlets); and 5) the root mats.
The upper 4 cm of the sandy stream bed was
sampled with 12 cylindrical cores (4.3 cm inner
diameter); the samples were fixed with formalin
(final concentration 2.5%) and the interstitial
fauna was elutriated using a 45-tpm mesh. Sam-
ples of detritus (approximately 2 cm3) were col-
lected from each area of the dark, fine deposits
in Cabaret stream and from bankside pools. Ob-
servations of macroscopic fauna in the cave
stream were noted during the whole study; 10
m of the open stream were also swept twice for
microscopic animals using a square (20 cm X
20 cm) net (250-1xm
mesh) with a collecting jar
attached to the distal, attenuated end of the net.
Root mat samples, typically 2-3 cm3, were col-
lected throughout
the study (including 10 Au-
gust 1990).
Samples
collected from all 5 habitats
were sorted live for animals using 160X mag-
nification.
Primary
function
of root mats:
food
or shelter?
To
discriminate
between the possible primary
function of the root mats-shelter or food-the
following experiment was carried out. On 23
May 1990, isolated branches of living root mat
were shaken free of animals and debris, and a
2 x 2 x 1-cm portion of root mat was excised
from each to verify (using 160x magnification)
that most of the root mat fauna had been re-
moved together with the debris. Within 0.5 m
of each branch
and under a similar stream flow
regime, a synthetic root mat was installed, of
similar dimensions, but made from inedible
split fibre of polyethylene
tape (an artificial
sea-
weed substrate with rough surfaces which
aquatic
microinvertebrates
can grip easily). Im-
mediately upstream of each pair of real and
synthetic
root mats a barrier
of 1x 1-mm mesh
was installed to prevent large clumps of root
mat, which were sometimes carried by the
stream, from settling on the substrates.
The 8
pairs of real and synthetic root mats were re-
moved from Cabaret stream
on 30 August 1990,
and the fauna of each was sorted live using
160x magnification.
Difference
in colonisation
of the 2 substrates was tested using a paired-
sample t test (1-tailed,
7 df). If the cavemicoles
feed mainly on some substance
in the water,
or
on microorganisms
present either in the water
column or settling on any available
substrate,
then the numbers of colonising animals should
be similar
for
both real and synthetic
root mats.
The faeces of several
specimens
of 2 common
crustaceans
from Cabaret
stream,
Austrochilton-
ia subtenuis
(Sayce)
1902
(Amphipoda:Ceinidae)
and the gilgie (a freshwater crayfish) Cherax
quinquecarinatus (Gray) 1845 (Decapoda:Paras-
tacidae), were examined using a compound
light microscope
at 400x magnification
to iden-
tify the foods consumed by these animals. Fae-
ces were collected
from a total of 20 amphipods
and 16 gilgies on 2 occasions in August 1990.
Amphipods (n = 40) and gilgies (n = 10) were
maintained under laboratory
conditions for 5
and 2 mo, respectively,
on a diet of pure tuart
mycorrhizae (harvested weekly). At various
times during the study, the behaviour
of some
1996] 513
This content downloaded from 142.244.226.141 on Thu, 3 Oct 2013 13:14:40 PM
All use subject to JSTOR Terms and Conditions
E. J. JASINSKA
ET AL.
invertebrates and the guts contents of a few oth-
er cave stream invertebrates with transparent
body walls (janirid isopods, chironomid larvae,
and Hurleya
sp.) were also observed using both
dissecting and compound microscopes and up
to 400x magnification.
Results
The physicochemical properties of Cabaret
stream were stable throughout the study with
very little variation in temperature: 19.0-19.3?C,
pH: 6.64-6.90, total alkalinity: 66-82 mgCaCO3/
L, dissolved oxygen: 5.6-7.1 mgO2/L, and con-
ductivity: 420-470 [S/cm.
Investigation of potential food sources
During the study, cadavers were not observed
in Cabaret stream nor collected on the mesh
barriers nor in sediment traps. Neither bats, nor
signs of their presence such as guano, were
sighted in the cave. The debris collected on
mesh barriers in shallow and deep sections of
the stream consisted only of rootlet fragments
and amorphous organic matter identical in col-
our and texture to that found deep in root-mat
interstices. In sediment traps, likewise, root-mat
debris and occasional fragments of exoskeleton
from cave crayfish were the only organic mate-
rials trapped. In parts of the stream there were
deposits of dark fine detritus overlying the
sandy streambed. The deposits of dark fine de-
tritus in slow-flowing sections of the stream
were identified as tannin-impregnated frag-
ments of rootlets. The organic matter separated
from core samples of the streambed also con-
sisted of root-mat detritus and sometimes of
considerable quantities of amorphous organic
matter (again indistinguishable from that found
in root-mat interstices).
Concentrations of dissolved amino acids, sug-
ars, and organic acids were all below the limits
of detection (0.01 nmol/L). Phenolic compounds
(most likely root tannins), however, were readi-
ly detectable by HPLC.
The root-mat branches in exclosure cages all
increased in area. The growth increments over
the 20 wk were in the range 30-400 cm2, rep-
resenting up to an 18-fold increase in root mat
area.
Microscopic study of the rootlets revealed
them to be mycorrhizal. Only some new growth
and thick mother roots were non-mycorrhizal.
Several different types of fungal hyphae iden-
tified as ascomycetes (characteristic spores),
zygomycetes (characteristic spores), and basid-
iomycetes (presence of clump connections on
hyphae) commonly occurred within the same
area of a rootlet. Some short lateral rootlets were
enclosed by fungal mantles while others were
penetrated intercellularly by hyphae and intra-
cellular fungal haustoria and, occasionally, fun-
gal spores. Free mycelial growth extending out
from the roots occurred only infrequently, and
then on scenescing root mats. All mycorrhizal
rootlets contained heavy deposits of tannins
both in the hyphae and in the outer cortical and
epidermal root cells.
Distribution of
fauna in stream microhabitats
The 12 cores from the upper 4 cm of the
sandy stream bed (50 g dry weight) yielded 7
species of microinvertebrates. Animal taxa were
generally represented by fewer than 10 animals
with the notable exception of nematodes. Al-
though some cores did not contain any nema-
todes, 2 cores with considerably more root-mat
detritus (including the amorphous organic mat-
ter) than the others had 185 and 193 nematodes
respectively. The streambed detritus harboured
just 2 species (nematodes and cyclopoid cope-
pods); the most fauna-rich sample contained
only 12 nematodes and 3 cyclopoid copepods.
The bankside pools contained eyeless amphi-
pods (Hurleya
sp.), 1 or 2 per sample, and nu-
merous rotifers (2 species). In the open water,
freshwater crayfish were commonly sighted,
and nightfish (Bostockia
porosa, Percichthyidae),
leeches (1 species), and dytiscid beetles (1 spe-
cies) were occasionally observed (sheltering un-
der the root mats when disturbed by torch
light). The 2 sweeps of Cabaret cave streamway
produced only 3 cyclopoid copepods. In con-
trast, the root mats yielded 23 species of inver-
tebrates (Table 2). Densities of root-mat fauna
varied from 50 to 600 animals per g dry weight
of root material. The numerically dominant tax-
onomic groups in the root mats included janirid
isopods, the amphipod Austrochiltonia
subtenuis,
the ostracod Gomphodella
?maia, and halacarid
mites.
[Volume 15
514
This content downloaded from 142.244.226.141 on Thu, 3 Oct 2013 13:14:40 PM
All use subject to JSTOR Terms and Conditions
1996] ROOT-MAT HABITATS IN CAVE
WATERS 515
TABLE 2. The microhabitat associations
of aquatic
fauna from Cabaret cave stream,
Yanchep
National Park.
Perthia
sp., trichopteran
larvae,
?Antipodrilus
sp., and Darwinula
sp. were only rarely
encountered.
Root Bankside Sand Open Streambed
Taxon mat pools interstices water detritus
Vertebrates
Osteichthyes:Percichthyidae
Bostockia
porosa /
Invertebrates
Acarina
Lobohalacarus
sp. nov. /
Soldanellonyx
sp. /
Oribatida
sp. 1 /
Oribatida
sp. 2 /
Oribatida
sp. 3 /
Crustacea
Amphipoda
Austrochiltonia subtenuis /
Hurleya
sp.
Perthia
sp. nov. /
Copepoda
Paracyclops sp. nov./ // /
Eucyclops
linderi
Harpacticoida sp. /
Decapoda:Parastacidae
Cherax
quinquecarinatus /
Isopoda
Janiridae
sp. nov. /
Ostracoda
Darwinula
sp. /
Gomphodella
?maia /
Syncarida
Bathynellidae sp.
Entognathous hexapod sp. /
Insecta
Diptera
Polypedilum sp. larvae /
Tipulidae
sp. larvae /
Coleoptera: Dytiscidae
Sternopriscus
sp./ /
Trichoptera
Leptoceridae sp. larvae /
Hirudinea
Erpobdellidae sp. /
Oligocheata
Aeolosomatidae
sp. /
Nais ?variabilis /
Pristina
longiseta /
Phreodrilidae
sp. /
?Antipodrilus
sp. /
This content downloaded from 142.244.226.141 on Thu, 3 Oct 2013 13:14:40 PM
All use subject to JSTOR Terms and Conditions
E. J. JASINSKA
ET AL.
TABLE
2. Continued.
Root Bankside Sand Open Streambed
Taxon mat pools interstices water detritus
Turbellaria
Stenostomum
sp. /
Catenulida
sp. /
Macrostomum
sp. 1 ,/
Macrostomum
sp. 2 /
Typhloplanoida
sp. /
Rotatoria
Bdelloidea
sp. / /
Ploima
sp. / /
Nematoda
6 spp. / / /
Primary
function of root mats:
food or shelter?
The removal of root-mat fauna from isolated
branches of root mat was effective because none
of the subsamples removed from each branch
contained any animals. By the end of the 3-mo
experiment, both synthetic mats and root mats
had been colonised to various degrees by the
aquatic invertebrates. An unexpectedly large
number of nematodes (too great to be counted
and far higher than encountered in undisturbed
Cabaret root mats) settled on both synthetic and
living root mats, especially on the larger mats,
which trapped greater quantities of amorphous
TABLE
3. Numbers of aquatic
invertebrates colon-
ising pairs
of root
mats and synthetic
mats (both
mats
in each pair trimmed to the same size) between 23
May 1990 and 30 August 1990. Nematodes settled in
great
numbers
(many
hundreds)
on all mats
both liv-
ing and synthetic; they were not counted and there-
fore are not included
in the table.
Total number Number of
of animals taxonomic
groups
Pair Synthetic Root Synthetic Root
No. mats mats mats mats
I 1 13 1 3
II 7 10 3 5
III 0 38 0 5
IV 3 352 3 10
V 7 64 5 4
VI 2 396 2 12
VII 5 55 4 8
VIII 34 91 5 9
organic detritus. Nematodes also were observed
in great numbers in sediment traps and cores
containing large amounts of amorphous organic
matter. Nematodes were excluded from statisti-
cal analysis of patterns of recolonisation because
1) more nematodes recolonised both substrates
than were originally present in undisturbed
Cabaret stream root mats 2) nematode abun-
dance was up to 19 fold greater than the total
abundance of other animal taxa that colonised
the substrates and 3) nematode abundance both
in the recolonised substrates and in the sand
cores increased in proportion to the quantity of
amorphous organic matter and thus had little to
do with the root mats being used as shelter or
food. Both the total number of animals and the
number of taxonomic groups were significantly
higher in the root mats than in the correspond-
ing synthetic mats (p < 0.05 for total numbers
of animals and p < 0.01 for the number of tax-
onomic groups) (Table 3). The only synthetic
mat with a large number of animals was from
pair VIII and it had trapped more root detritus
than any of the other synthetic mats (probably
due to its moderately large size and the local
flow regime).
The faeces of Austrochiltonia subtenuis consist-
ed exclusively of fragments of mycorrhizal root-
lets. Intact Cherax
quinquecarinatus
(gilgie) faeces
also contained mycorrhizal fragments and, to a
lesser extent, fragments of microinvertebrate ex-
oskeletons. All gilgies survived the 2 mo in the
laboratory on a diet of pure tuart mycorrhizae,
and the amphipods continued to reproduce
throughout the 5 mo. The gut contents of a few
516 [Volume 15
This content downloaded from 142.244.226.141 on Thu, 3 Oct 2013 13:14:40 PM
All use subject to JSTOR Terms and Conditions
ROOT-MAT
HABITATS
IN CAVE WATERS
other cave stream invertebrates,
including jan-
irid isopods, chironomid
larvae,
and Hurleya
sp.,
were visible through the body wall of the ani-
mals and consisted of root debris. Once the
peritrophic
membrane
enclosing the faecal pel-
lets was broken, the contents were indistin-
guishable
from fine root debris and amorphous
organic
matter
normally
found in the root mats.
Nematodes and acarines from several species
were observed with stylets embedded in the
rootlets, indicating that these animals are sap
feeders. The leech (Erpobdellidae),
the fish Bos-
tockia
porosa,
and the dytiscid beetle Sternopris-
cus sp. are all predators
of small invertebrates.
Also, the amphipod Perthia
sp. nov. was ob-
served capturing and feeding on smaller am-
phipods and janirid
isopods, while the pharynx
of Macrostomum turbellarians
(2 species) com-
monly contained one of the following:
rotifers,
harpacticoid copepods, early stage chironomid
larvae (Polypedilum sp.), or juvenile ostracods.
Some of the oligochaetes (not identified) con-
tained
fine root-mat
debris
in their
guts and oth-
ers were observed feeding inside decomposing
amphipods and isopods although it was not
clear
whether
they were feeding on the cadavers
or on the protozoans also present. The nema-
todes that occurred in high numbers where
there was abundant
amorphous organic
matter
(decomposing mycorrhizal
material
and faeces
of the cave animals)
probably
feed on these par-
ticulates and on bacteria which are likely to
grow on such substrates.
Discussion
The physicochemical stability of Cabaret
stream is consistent with the view that there
was negligible surface drainage entering the
system which could have carried
epigean debris
into the subterranean
stream.
During the course of this study, 29 new spe-
cies of aquatic
fauna
were recognised
in Cabaret
cave,
increasing
the total species number
for the
cave stream from 12 to 41, with many of the
species still undescribed. Some are apparently
endemic to the Yanchep
cave streams and pos-
sess distinct troglobitic
features such as lack of
body and eye pigmentation
and elongated an-
tennae (e.g.,
janirid
isopods and the amphipod
Hurleya
sp.). Others,
for example the amphipod
Austrochiltonia
subtenuis and the freshwater
crayfish Cherax
quinquecarinatus,
appear to be
relatively recent colonisers with widespread
conspecifics in surface waters. At least for A.
subtenuis,
no clear genetic separation
exists be-
tween cave and surface populations (Stuckey
1991),
though
compared
with surface
conspecif-
ics the cave individuals have much reduced
body pigmentation, fewer ommatidia, lower
number of eggs, greater antennal length, a
greater
number
of antennal
segments,
and high-
er egg volume (Burt
1982,
Stuckey
1991).
Thus,
it appears
that
Cabaret cave stream
is inhabited
both by recent colonisers
and relict
species.
Since there is no direct surface
drainage
into
the cave stream, the only way aquatic fauna
could enter the cave stream presently is by
swimming upstream through a series of col-
lapses from a surface pond about 100 m west
of the cave. The fully aquatic
Cabaret cave ani-
mals, therefore,
can be regarded as permanent
residents of this subterranean environment,
completing their life-cycle within the ground-
water stream. It is also possible that the terres-
trial insects with aquatic
larval stages may re-
main underground
even as winged adults.
The 29 new species in Cabaret stream were
recognised from several
microhabitats
in Caba-
ret stream,
but 22 (including
3 species of nem-
atodes with stylets) of the 41 total species were
found only in the root mats. Only 1 species,
the
bathynellid
syncarid
extracted
from
cores of the
sandy streambed,
was never found in the root
mat microhabitat. The species richness and
abundance of animals in the different stream
microhabitats
raises the question of the poten-
tial food sources in this groundwater
stream.
Surface organic debris percolating through
the cave ceiling may provide
a source of carbon
for the Cabaret stream fauna, but this contri-
bution would be restricted
only to the wet part
of the year since rainfall
in southwestern Aus-
tralia is highly seasonal. Nevertheless,
organic
matter filtering through the cave ceiling into
Cabaret
stream is not likely
to represent
a major
input of carbon,
since epigean debris was not
detected either on mesh barriers or in sediment
traps in the stream.
Nor is epigean debris
likely
to degrade
rapidly given the high phenolic
con-
tent of Australian
plants growing in the area.
Admittedly, dissolved sugars, organic acids,
and amino acids were measured
on 1 date only,
but it was during a rainy period and samples
were collected from both open water and from
root mat
interstices.
Dissolved amino
acids,
sug-
1996] 517
This content downloaded from 142.244.226.141 on Thu, 3 Oct 2013 13:14:40 PM
All use subject to JSTOR Terms and Conditions
E. J. JASINSKA
ET AL.
ars, and organic acids were not detected from
either open water or root mat interstices (<0.01
nmol/L). Root exudates are evidently seques-
tered rapidly, perhaps by the mycorrhizal fungi
and possibly by saprophytic fungi and bacteria
which are likely to be growing on the rootlets
and on root-mat detritus. Phenolic compounds
(most likely root tannins), however, were readi-
ly detectable by HPLC, but these structurally
complex compounds are a poor substrate for the
growth of microorganisms (Miinster and Chrost
1990). It is likely that fungi contribute nutrition-
al value to the tree rootlets, but the presence of
fungi also coincided with heavy tannin deposi-
tion and tannins can impede both digestion and
growth of bacteria (Howe and Westley 1988,
Miinster and Chrost 1990). Although we have
attempted to measure the abundance of bacteria
in the cave water, both from the open stream
and from the root mat interstices, and we have
observed bacteria (using epifluorescence mi-
croscopy) on root mat detritus and rootlets, we
do not present our data here because of diffi-
culties experienced with contamination of the
fluorescent dye used. The unpublished data we
do have, however, suggest very low abundance
of free bacteria. It is still relevant to measure the
abundance of bacteria attached to the rootlets,
organic debris, sand grains, and the air-water
interface to further elucidate the pathways of
carbon transfer in the cave stream. However, we
have no evidence of chemoautotrophic activi-
ty-such as emission of hydrogen sulphide or
formation of slimes, benthic or floating mats of
bacteria and fungi, such as have been observed
in the species-rich Movile Cave, Romania (Sarbu
et al. 1996)-from Cabaret cave. The sandy sub-
strate of the cave stream comprises only leached
quartz sand, root mat debris, and pieces of cal-
carenite; the Cabaret stream is, in effect, an
elongate spring issuing from the sandy substra-
tum. Therefore, even if moderately high num-
bers of attached bacteria are found, it is likely
that their primary carbon source is derived from
the roots directly (e.g., root exudates, scenescing
root cells) or indirectly (e.g., via the fungi, or
animals which derive their energy from the root
mats).
The growth of root mats measured outside
the maximum root-growth season (summer)
was still substantial and consistent with the idea
that the root mats provide a continuous and
abundant source of food for the cave stream fau-
na.
We conclude that the root-mat habitat consti-
tutes a discrete miniature ecosystem where the
primary source of energy is provided by the tree
roots and the fauna includes grazers, sap-feed-
ers, decomposers, and predators. With the pos-
sible exception of some of the sand-interstitial
invertebrates which are likely to be bacterivo-
rous, most of the Cabaret cave aquatic fauna ap-
pears to derive its source of energy directly or
indirectly from the mycorrhizal root mats.
Roots, often mycorrhizal, are common in
shallow groundwaters and hyporrheic zones of
wooded areas. The roots of reeds and sedges
are a widespread habitat in swamps and ripar-
ian zones. They represent a distinct habitat
which can provide not only shelter but also an
important food source where green food is lim-
ited or unavailable. Particularly, aquatic tree
roots are likely to be an important source of
food for cave and interstitial groundwater fau-
nas of shallow unconfined aquifers elsewhere.
Acknowledgements
We sincerely appreciate the field assistance of
many persons, all of whom are acknowledged
in Jasinska (1990). Special thanks are extended
to Department of Conservation and Land Man-
agement staff at Yanchep National Park, and to
T. Jasinski and Associated Surveys Pty. Ltd. for
their help in establishing benchmarks in Caba-
ret cave. We gratefully acknowledge funding
from the Endangered Species Program of the
Australian National Parks and Wildlife Service,
Murdoch University, and The University of
Western Australia. Voucher specimens are pres-
ently kept in the Department of Zoology, The
University of Western Australia, but on comple-
tion of the study, in 1997, will be transferred to
the Western Australian Museum.
Literature Cited
BARR, T. C. 1967. Observation on the ecology of caves.
The American Naturalist 101:475-491.
BURT, J. S. 1982. Aquatic cavernicoles of the Yanchep
Caves. Honours Thesis, The University of Western
Australia, Perth.
CULVER,
D. C. 1982. Cave life. Evolution and ecology.
Harvard University Press, Cambridge, Massachu-
setts.
EBERHARD,
S. M. 1992. The invertebrate cave fauna of
518 [Volume 15
This content downloaded from 142.244.226.141 on Thu, 3 Oct 2013 13:14:40 PM
All use subject to JSTOR Terms and Conditions
ROOT-MAT HABITATS IN CAVE WATERS
Tasmania: ecology and conservation biology.
Masters Thesis, University of Tasmania, Hobart.
HOWARTH,
F. G. 1972. Cavernicoles in lava tubes on
the island of Hawaii. Science 175:325-326.
HOWARTH,
E G. 1983. Ecology of cave arthropods.
Annual Review of Entomology 28:365-389.
HOWARTH,
E G. 1988. Environmental ecology of
north Queensland caves: or why there are so
many troglobites in Australia. Pages 76-84 in L.
Pearson (editor). Proceedings of 17th Biennial
Conference, TROPICON,
of the Australian Spele-
ological Federation, Cairns, Queensland.
HOWE,
H. F, AND L. C. WESTLEY. 1988. Ecological
relationships of plants and animals. (pages 31-
35). Oxford University Press, New York.
JASINSKA,
E. J. 1990. Root mat ecosystem in an epi-
phreatic stream; Cabaret cave, Yanchep National
Park. Honours Thesis, Murdoch University, Perth,
Western Australia.
JASINSKA,
E. J., AND B. KNOTT. 1991. Stability of root
mat ecosystems in a groundwater stream: Caba-
ret cave, Yanchep National Park, Western Austra-
lia. Australian National Parks and Wildlife Ser-
vice ESP Project No. 46., Canberra.
KENDRICK,
G. W., K. H. WYRWOLL,
AND B. J. SZABO.
1991. Pliocene-Pleistocene coastal events and his-
tory along the western margin of Australia. Qua-
ternary Science Reviews 10:419-439.
LONGLEY,
G. 1981. The Edwards Aquifer: Earth's
most diverse groundwater ecosystem? Interna-
tional Journal of Speleology 11:123-128.
MUNSTER, U., AND R. J. CHROST. 1990. Origin, com-
position and microbial utilization of dissolved or-
ganic matter. Pages 8-46 in J. Overbeck and R. J.
Chr6st (editors). Aquatic microbial ecology: bio-
chemical and molecular approaches. Springer-
Verlag, New York.
NOTENBOOM,
J. 1981. Amsterdam expedition to the
West Indian Islands, Report 12. The new hypo-
gean cirolanid isopod crustaceans from Haiti and
Mayaguana Bahamas. Bijdragen tot de Dierkunde
51:313-331.
REDDELL,
J. R. 1981. A review of the cavernicole fauna
of Mexico, Guatemala, and Belize. Texas Memo-
rial Museum, The University of Texas at Austin,
Bulletin No. 27.
SARBU,
S. M., T. C. KANE,
AND B. K. KINKLE 1996. A
chemoautotrophically based cave ecosystem. Sci-
ence 272:1953-1955.
SKET,
B. 1991. Isopoda (Crustacea: Isopoda:Micro-
charon,
Jaera,
Proasellus)
and other fauna in hypo-
gean waters of southern Cyprus. International
Journal of Speleology 19:39-50.
STUCKEY,
M. L. 1991. Differences between cave and
surface populations of Austrochiltonia subtenuis
(Sayce), 1902 (Amphipoda:Ceinidae)-selection
or developmental plasticity? Honours Thesis, Uni-
versity of Western Australia, Perth.
VANEK,
V. 1982. Fauna of groundwaters of Bohemian
karst (Barrandium). Methodology and prelimi-
nary results. Polskie Archiwum Hydrobiologii 29:
415-424.
Received: 24 May 1996
Accepted:
1 October 1996
1996] 519
This content downloaded from 142.244.226.141 on Thu, 3 Oct 2013 13:14:40 PM
All use subject to JSTOR Terms and Conditions