”3"“
'e.
I

 

, ,4. ——
V -w- -. .1. «gr.-. . .

EFFECTS OF DEVELOPMENTAL
STAGE AND PRIOR EXPERIENCE 0N.
HABITAT SELECTION m EHREE SPECIES
or mum LARVAE '

Thesis for the Degree of M. S.
MICHIGAN STATE UNIVERSITY
RiCHARD KNIGHT O’HARA
1974

......
......

v
x‘

' "‘RWWA w

\\ \

\ \

i

WM * V

  
 
  

LIBwlAPYl

I Mic ".h' gas 3:
UniveES.3r .y

"(MB 02680“?

 

‘ MK BINUERV EM. .

ABSTRACT

EFFECTS OF DEVELOPMENTAL STAGE AND PRIOR
EXPERIENCE ON HABITAT SELECTION IN
THREE SPECIES OF ANURAN LARVAE

BY

Richard Knight O'Hara

Three species of anuran larvae were reared in arti-
ficial habitats that varied only in visual appearance. At
different points in development, samples of single indi-
viduals from each habitat were allowed to choose between
two conditioning substrates to determine the influence of
developmental stage and prior experience on habitat selec-
tion. The design of the two standard test substrates
allowed discernment of the relative importance of midwater
and bottom cues in selection of a substrate. Results showed
highly stereotyped behavior dependent primarily on cues
situated in the vertical plane (midwater).

In a preliminary study, larvae of the wood frog,
Eggg_sylvatica, were reared from the hatchling stage in two
experimental habitats. Habitat A was composed of alter-
nating vertical black and white stripes affixed to the
sides and a uniform white bottom. Habitat B consisted of

identical black and white striping on the bottom and uniform

Richard Knight O'Hara

white sides. Following 20-25 days experience, when pre—
sented with a choice between Habitats A and B, larvae con-
ditioned in Habitat A (Group A) significantly selected
Habitat B in both initial choice and total time. Habitat B
conditioned larvae (Group B) also demonstrated preference
for Habitat B in distribution of total time. During testing,
spontaneous reactions by individuals of both groups when
Habitat A was approached indicated that this selection was
based on an avoidance to the vertical stripes of that
habitat.

Larvae of the American toad, §E£g_americanus, were
reared from the day of hatching through metamorphosis in
experimental Habitats A and B and a uniform white control
habitat (Habitat C). No experiental or developmental influ-
ence was indicated on days 9-11 and 23-25 posthatching when
larvae of all three groups demonstrated preference for
Habitat B in initial choice and total time. Swimming
behavior indicated strong avoidance to the vertical stripes
of Habitat A and activity for every group was much greater
while over the preferred substrate. This relation of
activity and habitat is discussed in reference to other
studies.

Tadpoles of Groups B and C were again tested for
preference on days 34-37 when they exhibited varying stages
of transformation. Larvae of both groups continued to
avoid Habitat A in initial choice and total time. However,

those with tails resorbed to 7 mm or less reversed choice

Richard Knight O'Hara

behavior and strongly selected Habitat A. This sudden
developmental change seems to represent strong attending to
cues resembling a climbable surface and is substantiated by
swimming behavior within Habitat A during testing.

When E, americanus larvae were presented with a

 

choice between Habitat B and Habitat C on days 32-33, groups
tended to prefer Habitat C. This substantiated indications
that the side striping of Habitat A was the avoided com-
ponent and suggested that the bottom striping was not only
discriminated, but was also aversive in a lesser degree.
Larvae of the green frog, 5223 clamitans, were
reared from hatching in the three standard conditioning
habitats and tested for preference throughout development.
On days 10-11, all groups significantly chose Habitat B in
initial choice. Group A tadpoles continued to demonstrate
preference for Habitat B on days 25-28, however, Groups B
and C exhibited no preference. It appears that a general
developmental change in preference occurred at this stage
and prior exposure of Group A to Habitat A was sufficient to
overcome this change. On days 52-54, Groups A and B tended
toward selection of Habitat B, but not significantly.
Group C strongly preferred Habitat B in this test and again
on day 66. The final standard test on days 81-83 post-
hatching showed Groups B and C avoiding Habitat A while
Group A indicated an influence on selection attributable to

prior conditioning by exhibiting no preference.

Richard Knight O'Hara

A variety of novel substrates, substituted in place

of Habitat A, were tested with 5. clamitans in an effort to

 

determine the "aversive" components of the vertical stripes.
Aversive visual flicker was eliminated as a factor by
testing with horizontal stripes. Overall reflective inten-
sity of the substrate was likewise discarded as a single
influencer of selection by testing uniform gray and uniform
black sides. Tadpoles did not actively discriminate
vertical segments of aquatic vegetation (Elodea canadensis)
and when horizontal "fringe" was added to the black stripes
of Habitat A, tadpoles continued to strongly avoid that
substrate. By manipulation of widths and ratios of the
vertical black and white stripes, it was concluded that the
ratio of the contrasting stripes was important in eliciting
avoidance to Habitat A in this species. Adaptive impli-

cations of these behaviors are discussed.

EFFECTS OF DEVELOPMENTAL STAGE AND PRIOR
EXPERIENCE ON HABITAT SELECTION IN

THREE SPECIES OF ANURAN LARVAE

BY

Richard Knight O'Hara

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Zoology

1974

\3 ACKNOWLEDGMENTS
{{x
k9 .

I greatly appreciate the guidance and suggestions
provided by my major professor, Dr. Marvin M. Hensley,
toward completion of this thesis. I also extend thanks to
Dr. Eugene W. Roelofs and Dr. John A. King for serving on
my committee. Dr. Roelofs not only offered pertinent advice
during this investigation, but in conjunction with the
Fisheries and Wildlife Department secured the invaluable
laboratory space for this project. Dr. King provided many
helpful ideas and was a continual inspiration through our
discussions and his critique. I am grateful for the special
assistance furnished by the Department of Zoology in
acquiring experimental materials.

Dr. Robert L. Raisler of the Department of Psy-
chology supplied useful comments on statistical analysis.
The help of Mr. James Lucas in the construction of the
experimental tanks is also appreciated.

Among the many people providing various suggestions
and consultation, special thanks go to John Fitch, Chris

Thoms, and Dr. Martin Balaban.

ii

TABLE OF

LIST OF TABLES . . . . .
LIST OF FIGURES. . . . .
INTRODUCTION. . . . . .
METHODS AND MATERIALS. . .

Species Selection . .

Choice of Experimental Visual

Conditioning and Testing .

Rana sylvatica. . . .
Bufo americanus . . .
Rana clamitans. . . .

Test Apparatus and Procedure. ., . . . . . .
Statistical Treatment of Data

RESULTS 0 O O I O O 0

Testing of Rana gylvatica.
Testing of Bufo americanus

Behavioral Observations.

Testing of Rana clamitans.

CONTENTS

Page

vii

13

13
14
15

Parameters. . . .

15
16
17

19
21

22

22
25

32
35

Testing of Conditioned Substrates:

Developmental and Experiential Effects . . .
Investigation of the Aversive Component . . .

DISCUSSION . . . . . .

Developmental Affects . .
Experiential Affects . .

Selection Based on Avoidance. . . . . . . .

36
39

46

46

50

iii

Page

Activity Over Test substrates . . . . . . . 57
Relative Importance of Midwater and Bottom
Stripes I I O O O O O O O I O O O O 59
SUWARY O O O O O O O O O O O O O O O 6 1
LITERATURE CITED . . . . . . . . . . . . 66

iv

Table

1.

LIST OF TABLES

Mean data distribution of initial choices,
total time, activity, and side changes in
standard testing of B. americanus
tadpoles . . . . . . . . . . . .

 

Probability values of initial choice distri-
butions (chi-square test) and percent
total time spent over respective sub-
strates (t-test) by B, americanus larvae .

One-way analysis of variance of percent
total time spent in Side A byB
americanus during Test I. . . . . . .

One-way analysis of variance of percent
total time spent in Side A by g
americanus during Test II . . . . . .

One-way analysis of variance of percent total
time spent in Side A for B. americanus
tadpoles with tail lengths greater than
7 mm and less than or equal to 7 mm in
Test IV . . . . . . . . . . . .

 

Two-way anova with replication of the percent
total time spent active by B. americanus
over Habitats A and B in Test III. . . .

Two-way anova with replication of the percent
total time spent active by B. americanus
over Habitats A and B in Test II. . . .

Mean data distribution of initial choices,
total time, activity, and side changes by
a. americanus between Habitats B and C
in Test III . . . . . . . . . . .

Page

26

27

30

30

31

31

32

33

Table Page

9. One-way analysis of variance of percent total
time spent in Side C by g, americanus
during Test III . . . . . . . . . . 34

10. Summary of standard test results of R.

clamitans showing distribution of Initial
choices and prerelease orientations and
their probabilities. . . . . . . . . 37

11. Summary of tests investigating "aversiveness"

by R. clamitans showing initial choices and
theIr probaBilities. . . . . . . . . 40

vi

LIST OF FIGURES

Figure Page
1. Experimental test tank showing Habitats A

and B O O O O O O O O O O O O O 2 0

2. Test substrate used in Habitat D. . . . . 44

vii

INTRODUCTION

Within the geographic range of a species, animals
are neither uniformly nor randomly distributed. Rather,
their spatial patterning is associated with particular
features of the immediate environment, the bounds of which
are largely dependent upon interactions with other indi-
viduals and species and certain environmental constraints.
Factors such as competition and predation as well as a
species' physical limitations and physiological tolerances
are recognized as primary influences acting to restrict
animals to defineable habitats. Until relatively recently
(Lack, 1933), little attention has been directed at possible
psychological influences under which single or varied
environmental parameters are responded to preferentially.
Habitat selection as such implies discrimination and active
preference and should not be confused with habitat corre-
lation (Klopfer, 1969).

Considerable evidence offered in support of habitat
selection in vertebrates is of a correlative and indirect
nature. There is difficulty in directly dealing with a

problem of such complexity which necessarily demands

simplistic procedural analysis in demonstrating its actu-
ality and defining its primary components. Experimental
procedure in standard design allows an animal to actively
choose between two or more "habitats" differing in one or
more characteristics. The selected habitat is designated
as the one initially selected or the one in which the
majority of time is spent.

Experimental studies of this type conducted in
highly simplified artificial conditions yield questionable
results in terms of relevance to the natural situation.
However, such an approach can enable an experimenter to
begin to assess an animal's discriminative capabilities,
the class and components of conditions preferred or avoided,
and the underlying basis for the selective process itself.
Coupled with field observations, predictions can be made
about an animal's behavior in certain natural circumstances
and ideas formulated concerning its adaptiveness.

Much of the recent work investigating habitat
selection is with fish and indicates a variety of potential
parameters affecting preference. Sale (1968 and 1969a)
reports several proximate factors influencing habitat
selection in the juvenile manini, Acanthurus triostegus.
The presence of substratum, cover, food algae, and the depth
of the water were all of importance. Plant morphology was
apparently a strong determinant of distribution within a
stream for fourspine sticklebacks (Baker, 1971). Hagen

(1967) investigated mechanisms acting to isolate two

subspecies of the threespine stickleback, Gasterosteus

 

aculeatus, in a stream. Results of field observations and
preference tests showed distinct differences in preference
of type of nesting substrate, nesting vegetation type, and
selection of water current. Other investigations indicate
that factors such as overhead cover, light intensity,
shelter complexity, water velocity, depth, and turbulence
may be important cues to some fish in selecting micro-
habitats (Hunsaker & Crawford, 1964; Kwain & McCrimmon,
1967; McCrimmon & Kwain, 1966; Butler & Hawthorne, 1968;
Haines & Butler, 1969; Baldes & Vincent, 1969; and Sale,
1972).

Hypothetical mechanisms of habitat selection and
their ontogeny have been discussed in reference to fish
(Sale, 1969b) and birds (Svardson, 1949; Hilden, 1965; and
Klopfer & Hailman, 1965) but offer little substantive
experimental data. Preliminary work has sought to demon-
strate the reality of a selective process and define the
variables capable of modifying and structuring it. Testing
of the underlying basis of habitat selection has been
incorporated into several recent studies. It appears that
learning may play a significant role.

Sale (1971) presented representatives of two popu-
lations of the reef fish, Dascyllus aruanus, with a choice
between two species of coral that each group was found to
frequent and a third "less desirable" species. Fish

selected to occupy the coral from which they were collected,

indicating that prior experience may have affected subse-
quent preference. Early experience within the first 60
days of life was a necessary prerequisite for subsequent
selection of a stone or sand substrate in the African
cichlid, Tilapia mossambica (Quertermus, 1972), but does
not influence nest-site selection which is based on the
substrate of smallest particle size (Thoms, 1974).

Klopfer (1963, 1965) and Klopfer and Hailman (1965)
studied experiential influences on foliage preference in
small numbers of chipping sparrows. Although choices
seemed primarily genetically dependent, there were indi-
cations that learning through early experience may have
modified later foliage preference. Sargent (1965) showed
that the conditions under which young zebra finches were
reared affected selection of nestling materials and habitat
in mature individuals.

Studies with mammals have indicated a strong
hereditary basis for selection of a habitat. In field and

woodland subspecies of the deermouse, Peromyscus maniculatus,

 

Harris (1952) found that mice preferred habitats resembling
their natural environment even without prior outdoor experi-
ence. Harris concluded that this behavior was innate,
although possible experiential influence was untested.
Wecker (1963) studied this problem in greater depth with

the prairie deermouse, g. m. bairdi. Habitat choice of its
natural field habitat was primarily hereditary as early

experience in a woods or laboratory habitat did not modify

preference for the field habitat. Animals reared for 12-20
generations in thelaboratory lost their genetically based
preference, but retained the ability to reestablish prefer-
ence following experience in a field situation.

Psychological preference for a habitat seems a
profound influencer of microdistributional patterns in many
vertebrates. There is, however, little basis for such an
assertion in regard to the amphibians. Supportive evidence
of habitat selection is mostly correlative and lacking in
experimental data.

In review of the literature, it seems that any
selection process occurring in adult amphibians is most
probably dependent on critical moisture and temperature
levels or in locating suitable breeding and nesting sites.
Heatwole (1962) reports that the microdistribution of the

red-backed salamander, Plethodon cinereus, is dependent on

 

temperature, humidity, and log and litter type. The narrow-
mouthed toad, Microhyla carolinensis, is limited by moisture
and cover, with distribution and density correlating with
plant associations and the depth of penetrable organic
layers (Anderson, 1954). Ralin (1968) found indications
that microclimatic differences in relative humidity was
sufficient to separate two sympatric species of hylids
ecologically and reproductively. Moisture also seems to
govern the distribution of the cave salamander within the

twilight zones of limestone caves (Hutchison, 1958).

Heatwole (1961) investigated local distribution

patterns of the wood frog, Rana sylvatica, in southern

 

Michigan. Locations were restricted almost entirely to
moist, forested areas. Arborescent vegetation providing
shade in conjunction with a nearby body of water seemed
important. When tested experimentally, selection of litter
structure was influenced by substrate moisture.

The influence of temperature on substrate selection
in different deve10pmental stages of the clouded salamander,
Aneides ferreus, was studied by McKenzie and Storm (1970).
Young salamanders selected bark litter over all temperatures,
subadults selected bark litter only at higher temperatures,
and adults exhibited no preference between bark litter or a
rocky substrate.

The Specificity with which breeding and oviposition
sites are selected, particularly in the anurans, has been
commonly observed. Breeding site locations and oviposition
sites are often correlated with such factors as the size,
permanence, and quality of a body of water, the character-
istics of the surrounding terrain, aquatic vegetation, and
water depth. Bragg (1940a) and Bragg and Smith (1942) when
studying the ecology of §u£2_cognatus, noted breeding to be
confined to clear, shallow water. While calling males were
indiscriminate in choice of a mating location, females were
very selective in choosing larger, longer-lived pools.
Breeding, nest site selection, and local segragationing of

overlapping breeding Species are reviewed for many anurans

of Oklahoma by Bragg (1940b) and Bragg and Smith (1942).
Interesting to note is the observation by these authors

that the toad Bufo woodhousei usually calls while sitting

 

in the water, whereas B. americanus when breeding at the
same time will call from the banks. The opposite condition
was observed by Brown and Pierce (1965) in southern Illi-
nois.

A few studies report greater specificity of nest
site selection than correlation with the gross features of
water-type and location. Field observations suggest that

the eggs of the northern chorus frog, Pseudacris triseriata,

 

are laid on slanting vegetation (Wells, 1924) while 2:
clarkii lays its eggs on upright vegetation (Bragg, 1943).
Sexton and Ortleb (1966) cite experimental evidence

indicating that the leptodactylid frog, Engystomops

 

postulosus, responds to shallow water and a curved surface
or a solid vertical plane intersecting and perpendicular to
the surface of the water.

Perhaps the least understood period in the life of
amphibians is the larval stage. This is, however, where
the greatest mortalities are sustained and primarily where
the size of the subsequent breeding population is ultimately
determined. Predation on larvae is heavy, and severe inter-
and intraspecific competition for critical factors such as
food and oxygen and the process of growth inhibition can
become critical at high densities, retarding development

and indirectly or directly causing mortality (Savage, 1952;

Brockelman, 1969; Wilbur, 1972; and Wilbur & Collins, 1973).
A possible method for lessening the detrimental effects of
such interactions might be development of a microhabitat
preference. This would certainly be advantageous in cir—
cumstances where two or more species are developing in the
same body of water and could function to decrease inter-
actions with predators.

Microdistributional data and information about
ecological relationships of anuran larvae are severely
lacking. Bragg (1940a) observed that tadpoles of Bufo
cognatus were restricted entirely to shallow waters at one
end of a pond where breeding in the adults was previously
limited. Larvae of B. canorus appeared to congregate in
warm shallow waters of a pond during the day and deeper
waters at night (Mullally, 1953). Brockelman (1969) reports

the distribution of B. americanus tadpoles to be restricted

 

within 2.5 meters of the water's edge in lakes of southern
Michigan but to be unrestricted in marsh habitats. Most of
the somewhat rare ecological observations of anuran larvae
are of this incidental type, with little concern as to the
variables that may take part in localizing different
species.

A more thorough study was accomplished by Heyer
(1973) on several sympatric species of anuran larvae in
Thailand. Relative environmental segregation was main-
tained not only through utilization of different ponds, but

also by general spatial partitioning of the living space

within a pond. However, possible variables governing this
distribution were not discussed.

Temperature can be a very important factor to
amphibian larvae in lessening time to transformation, often
having substantial influence on subsequent population size.
Quite specific temperature selection by anuran larvae in the
laboratory and the field has been demonstrated repeatedly
(Workman & Fischer, 1941; Brattstrom, 1962; Herreid &
Kinney, 1967; Lucas & Reynolds, 1967; and Vlaming & Bury,
1970).

Substrate type selection by the tailed frog,
Ascaphus Erugi, was investigated experimentally by Altig
and Brodie (1972). When tested in an artificial stream,
tadpoles preferred substrates of smooth stones of specific
diameter. Selection of such a substrate in fast-flowing
streams is probably not entirely of psychological choice
but seems related to the structure and size of the oral
disc. Tadpoles were also positively rheotactic and nega-
tively phototactic.

Few studies have attempted to examine learning
ability in larval amphibians, and this potential is appar-
ently regarded with pessimism. Munn (1940) demonstrated

that small numbers of Hyla cinerea, H. versicolor, and Rana

 

catesbeiana were capable of learning Y and T mazes. Also,
tadpoles of Hyla arborea performed simplistic tasks of form

discrimination (Pache, 1932).

10

Until recently, no substantive experimental data
showing ability of anuran larvae to visually perceive and
prefer particular habitats nor its underlying basis was
available. Preliminary study into these questions was
initiated by Wiens (1970 and 1972) using two species of
Rana conditioned in uniform striped, square-patterned, and
featureless substrates. Tadpoles of the red-legged frog,
R. aurora, reared in featureless or square-patterned
habitats showed no preference at any time for either striped
or square-patterned substrate while stripe-reared larvae
significantly preferred the striped habitat (Wiens, 1970).
This preference, established within the first two weeks of
conditioning was retained after isolation from the con-
ditioning substrate and could readily be renewed in young
and old tadpoles. In a similar experiment (Wiens, 1972),
R. cascadae tadpoles demonstrated selection only following
conditioning in the square-patterned substrate. Obser-
vations of the developmental ecology of these species
indicate that these learned preferences may reflect the
natural larval habitat.

The present study was undertaken on several bases.
There exists considerable evidence supporting habitat dis-
crimination and selection of breeding and oviposition sites
in adult anurans. Some anuran larvae are capable of dis-
cerning and selecting certain environmental characteristics
(e.g., temperature) and have been observed to exhibit a

limited degree of Species specific Spatial segregation.

11

The ability to learn simple visual cues has been demon-
strated in some tadpoles and experientially—dependent sub-
strate preference was Shown to occur in two species of frog
larvae. Indeed, habitat selection of a learned or genetic
basis could serve as an important mechanism in lessening
pressures arising from predation or competitive interaction
or assuring rapid, healthy development.

It was the design of this investigation to rear
anuran larvae in three artificial habitats differing only
in visual characteristics and to test subsequent preference
of two of these habitats at the earliest possible date and
throughout development. In addition to observation of
developmental and experiential influences, it was hoped some
insight into the visual components of the habitats respon-
sible for any stereotyped choice pattern could be deduced
in addition to the relative importance of midwater and
bottom cues.

Three local Species of anurans were chosen for this
study. Tadpoles of the wood frog, Raga_sylvatica, were
used in a preliminary experiment to determine the effective-
ness of the chosen visual parameters in producing a
response. Wood frogs are found in relatively mesic forests
in close proximity to water. Breeding occurs in late March
through April in transient woodland pools were eggs are
attached to vegetation or submerged objects in Spherical

masses. Transformation is completed in six to 12 weeks.

12'

A second series of experiments employed larvae of
the American toad, Bufg americanus, and investigated
experiential and developmental influences on choice be-
havior using visual parameters identical to those of the
previous test. Adults of this species are found in more
xeric habitats in grasslands, sparse woods, and frequently
in gardens and farmland areas. Breeding is from late March
or April and may extend as late as July, with peak breeding
in mid-April. Larvae normally metamorphose in six to 10
weeks.

Tadpoles of the green frog, Rang clamitan§,_were
used in a final series of tests. These were subjected to
the same rearing experiences and in addition to testing for
experiential and developmental affects, the pertinent visual
component of the experimental substrate was investigated.

5. clamitans inhabit swamps, larger ponds, and sometimes

 

streams, always in close association with water. Eggs are
laid from May into August in large flattened masses in the
surface film amid emergent and floating vegetation. Eggs

laid early in the season may transform by early fall while

others overwinter and metamorphose the following year.

METHODS AND MATERIALS

Species Selection

 

Choice of anuran Species for experimentation en-
tailed three basic determinants: (1) the ease with which
eggs could be obtained at the time of an experimental
series; (2) the degree to*which each species was related to
the other; and (3) the type habitat in which each species,
adults and larvae, is frequently found.

Rana sylvatica is among the first frogs to emerge

 

from hibernation and to commence breeding. These not only
provided an excellent opportunity for early testing of
experimental design and parameters, but offered an inter-
esting alternative in habitat type. Larval development most
frequently takes place in shallow, transient, woodland

pools quite devoid of any substantial vegetative growth.

Bufo americanus, the most xeric of the forms tested,

 

begins breeding quite early and may continue through June
or July. Choice of oviposition sites is somewhat indis-
criminate, thus exposing larvae to varied aquatic habitats.
Development usually extends early into the summer season
when pond characteristics (e.g., vegetation) have changed

from those present in the spring.

13

14

Rana clamitans breed throughout most of the summer

 

months in the more permanent, swampy areas. Larvae may
overwinter and are consequently subjected to nearly all

stages of seasonal pond succession.

Choice of Experimental Visual Parameters

In selecting a visual parameter, the primary
objective was not to reproduce a natural component of the
larval aquatic habitat. But rather to create a readily
distinguishable habitat, simplistic in design, that would
test the comparative relevance of midwater versus bottom
cues. However, in view of Wien's (1970) results, suggesting
that a striped pattern may have resembled linear vegetation
to Raga aurora tadpoles, stripes were chosen as the experi-
mental pattern for the present study. Indeed, one of the
only obvious visual characteristics that differs within and
between southern Michigan ponds is vegetation type and
density.

Experimental Habitat A was composed of a white,
featureless bottom and alternate one-half inch wide black
and one-half inch wide white vertical stripes fastened to
all Sides of the tank. Habitat B consisted of identical
striping on the bottom of the tank and white Sides. Con—
trol Habitat C was composed of uniform white sides and

bottom.

15
Conditioning and Testing

Ragglsylvatica

On 31 March 1973, several hundred eggs from multiple
clutches of the wood frog, 5, sylvatica, were collected in
a woodland pool in Meridian Township, Michigan, approxi-
mately one—fourth mile north of M-43 and one-fourth mile
west of Meridian Road. These animals were to be used in a
preliminary study to test methods and to determine if the
chosen visual parameters could be discriminated and differ-
entially treated.

Hatching commenced and was completed on 4 April, and
approximately 50 larvae were transferred to each of two
glass aquaria filled to a depth of 20 cm (26 liters) with
dechlorinated tap water. Aeration was provided by air
stones. To the base and Sides of each tank was affixed
adhesive, white plastic vinyl to the height of the water-
line. In rearing tank A, one-half inch wide black vinyl
strips were attached vertically to all sides at one-half
inch intervals. Stripes of the same width and interval were
positioned widthwise on the base of rearing tank B. Larvae
were fed rabbit pellets or "TetraMin Staple Tablet Food"
twice daily and water was changed every other day. Photo-
period was uncontrolled and temperature fluctuated between
20 and 23°C. On days 20-25 posthatching, larvae were

tested for preference.

16

B222 americanus

On the night of 10 May, 1973, one clutch of approxi-
mately 5,000 eggs was collected from the west end of a
permanent pond located in a field immediately south of the
Natural Resources Building at Michigan State University,
East Lansing. Until hatching, eggs were placed in one
standard control rearing tank. All conditioning tanks were
structured of three-fourth inch marine plywood, sealed with
gray epoxy paint, and measured 87.5 x 57.0 x 30.5 cm. Each
was uniformly lined with white plastic sheeting. Tanks
were filled with charcoal filtered water (100 liters) to the
height of the plastic Sheeting (20 cm) and aerated with
Single air stones. Black vinyl Strips were affixed to
Habitats A and B as previously described.

Hatching was completed on 13 May, when larvae were
transferred to respective rearing habitats. These con-
sisted of one tank each of types A and B and three feature-
less, white control tanks (C). Ninety larvae were placed
in each habitat. All but one control tank were positioned
on a stand directly beneath a uniform double row of 40W cool
white lights. Lights were suspended 60 cm above the water
surface. The remaining control tank was situated on a
stand 61 cm below the other tanks. This third control tank
was employed as a possible reserve resource of animals and
for possible use in subjecting larvae to subsequent periods

of varied experiential treatment or retention tests.

17

All conditioning tanks were placed on a controlled
light regime of 16L:BD (on at 800 hours and off at 2400
hours). Laboratory temperatures normally fluctuated between
21 and 23°C but reached as high as 26°C during the last 10
days of the experiment. Tadpoles were fed "TetraMin Staple
Tablet Food" twice daily and all fecal material, uneaten
food, and most water was siphoned out of every rearing tank
at two day intervals.

A total of four tests were initiated on days 9, 23,
32, and 34 posthatching with each test period spanning 2-3
days. In all tests except Test III on day 32, individuals
were presented with a choice between habitat types A and B
(also designated Side A and Side B) to determine develop-
mental and experiential effects. Test III substituted
Side A with an all white control habitat. Experimentation

was terminated at metamorphosis.

Rana clamitans

 

On 4 July, one clutch of recently deposited eggs of
the green frog, 5, clamitans, was collected in the north-
west corner of a marshpond in the Rose Lake Wildlife Re-
search Area, 12 miles northeast of Lansing, Michigan. This
pond was located in the center of the Orange Area (between
Woodbury and Peacock Roads) just southwest of the region of
flooding. One-hundred twenty eggs were immediately placed
in each of the five rearing tanks previously described.

Hatching occurred on 6 July, and larvae were henceforth

18

reared under the same procedures as employed with larvae of
E. americanus.

To determine if conditioning and/or developmental
stage had any influence on habitat choice, all three groups
were tested for preference of Habitats A and B on days 10,
25, 52, and 81 posthatching. A standard test was also
carried out on day 66 with Habitat C reared larvae for
indications of possible change in behavior during that time
interval.

In view of results obtained with R, sylvatica and

 

B, americanus, it was decided to increase numerically the
larvae in each conditioning habitat to enable investigation
of the primary component of the visual parameter employed
that was effecting observed responses. During time inter-
vals between standard tests, several tests were performed
varying the visual characteristics of the habitats. These
comprised altering widths and proportions of stripes of
Habitat A, changing pattern, and testing the affects of live
aquatic vegetation (Elodea canadensis) on habitat selection.
In the later test, vertical segments of Elodea extending to
the water surface were arranged in three staggered rows in
two-thirds of one side of the test tank. Rows were 5 cm
apart and individual segments were placed at 10 cm inter-
vals within each row. A second test approximately doubled
this density. Experimentation was terminated on day 96

posthatching when numbers were depleted.

19

Test Apparatus and Procedure

The test tank (Figure l) measuring 57 x 57 x 30.5
cm was divided equally into the two rearing substrates, A
and B, and filled with 65 liters of charcoal filtered water
to the depth of the substrate (20 cm). One light unit con—
taining two 40W cool white lights was centered 78 cm
directly above the water surface of the tank. Lights and
test arena were completely enclosed with a dark, opaque
curtain with a single, narrow slit through which obser-
vations were made.

Standard testing procedure was established follow—
ing experimentation with methods during preliminary testing

of R, sylvatica. In testing of B, americanus and R,

 

clamitans larvae were randomly chosen for testing in
sequential series from respective habitats to avoid clumping
of individuals of one rearing habitat at any particular time
period or day of testing. Tadpoles were placed singly into
a 10 cm diameter plexiglas cylinder, open at both ends, and
situated vertically in the center of the test tank. After

a given time period (10 minutes) and when the tadpole was
motionless on the bottom, orientation within the release
chamber was recorded. The release chamber was then lifted
to the top of the enclosure by external means and behavior
of the tadpole was observed for a 10 minute period. The
Side to which initial movement was directed and into which
the tadpole penetrated approximately 6-8 cm or beyond was

defined as an initial habitat choice. Total time and

20

 

||\\\\

 

Figure 1. Experimental test tank Showing Habitats A
and B.

21

activity Spent in each habitat were measured using two
”GraLab" timers and one stop watch. The number and direc-
tion of Side changes and comments regarding general behavior
were noted. Following testing the total length of each
individual was recorded before discarding from the experi-
ment. At the end of each day of testing, the sides within
the test tank were reversed to control for the influence of

cues internal to the testing apparatus.

Statistical Treatment of Data

Four measurements used to indicate habitat prefer—
ence were subject to statistical analysis. Goodness of fit
for distributions of pre-release orientations and initial
choices were analyzed using a chi—square test. Variation
from 50 percent within groups of percent total time spent
over one substrate were analyzed by use of a t-test (Sokal &
Rohlf, 1969). A single classification analysis of variance
was calculated on total times Spent per substrate between
groups within a test. To determine if activity times
varied with experience or test substrate, one factor two-
way analyses of variance with replication (Sokal & Rohlf,
1969) were run on the percent of the total time over each

habitat that was spent in activity.

RESULTS

Testing of Rana sylvatica

Larvae of 3223 Sylvatica were conditioned in
respective habitats (SO/tank) and tested for preference on
days 20-25 posthatching. Substantial mortalities occurred
during rearing in each group. The cause for the mortality
is unknown, however, largest numbers were lost immediately
following water changes suggesting that either temperature
or mechanical disturbance was responsible. In total, 33
larvae conditioned in a habitat of striped sides and
featureless base (Group A) and 18 conditioned in a habitat
of featureless Sides and striped base (Group B) were tested
in a test tank bearing equal proportions of each substrate.

Larvae of Group A were tested on days 20-24 post-
hatching. Average total length at this period was 27.2 mm.
Orientation within the release chamber immediately prior to
release did not vary between sides as 12 of the tadpoles
oriented toward Substrate A and 13 toward Substrate B.
However, only 7 of 25 larvae tested initially selected Sub-
strate A. This was significant at the 5 percent level when

subjected to a chi-square test.

22

23

Following periods of 10-20 minutes within the
release chamber, individuals A1 through A10 were released
and observed for 12 minutes and larvae All through A25 were
observed for 15 minutes. (There were no observable differ-
ences in measured parameters due to varied test period
lengths.) Larvae spent on an average 28 percent of their
time within Habitat A and 72 percent within Habitat B.
Individual values over each substrate varied significantly
from random (p < .001) when subjected to a t-test.

Time of activity per Side was recorded for indi-
viduals All through A25. Expressed as a percentage of
respective total times, equal values of 21 percent were
calculated as time spent swimming over each substrate.

Side changes ranged from 0 to 8 with a mean of 2.

AS an additional test of the strength of initial
choice as an indicator of preference, four individuals were
released from the center of Habitat A and four from the
center of Habitat B (A26-A33). Three of the tadpoles
released from Habitat A initially selected Habitat B and
all four released from Habitat B remained in that side.

A mean of 25.2 mm was recorded for total length of
Group B larvae. Of 18 tested, nine were released from the
center of the test tank and nine from the center of Habitat
A. In neither case did larvae Show significant orientation
prior to release. Seven released from the center of the
test tank initially chose Habitat B. This was not signifi-

cant; however, percent total times of 8.4 and 91.6 for

24

Sides A and B respectively were highly Significant (p <
.001). Percent active time for this group was 19.7 percent
in Side A and 13.2 percent in Side B. Side changes ranged
from 0 to 4 with a mean of 2.

No significant differences were noted in any cate-
gory for nine larvae released within Habitat A. Six ini-
tially chose Side B and average total times of 44.9 percent
and 55.1 percent were apportioned to Sides A and B. Only
5.2 percent and 9.7 percent comprised active movement and
overall side changes decreased to one.

Behavioral observations recorded for each indi-
vidual tested support the indication that Habitat B was the
preferred substrate of both groups and suggest that larvae
may have actually been avoiding the habitat with striped
sides. It was commonly observed in both groups that tad-
poles approaching Side A from Side B would either quickly
reverse their direction of movement upon reaching the
border, actually enter Side A for a very short distance
before returning to Side B, or enter Side A and stop within
4-5 cm of the border and remain motionless on the bottom
before ultimately returning to Side B.

Although unreflected in orientation prior to
release, a considerable proportion of tadpoles, predomi-
nately of Group B, exhibited persistent swimming movements
toward Side B during the period Spent within the release

chamber.

25

Testing of Bufo americanus
Because of clear discrimination and selective be-
havior demonstrated by R. sylvatica, identical visual cues

were retained for controlled testing of Bufo americanus

 

larvae.

Through development, three standard tests of prefer-
ence of Substrates A and B were conducted. Summary of these
data is given in Table l, and statistical treatment of
initial choices and total times is presented in Table 2.
Total time over each test substrate is given as a percent-
age of a ten-minute test period. Activity times are in
percents of total time over that substrate. Pre-release
orientation was randomly distributed for every group in
each test and is not included in data summary.

Test I was initiated at 9 days posthatching at
stages 24-25 (Gosner, 1960) when larvae began to show
coordinated swimming movements, feeding activity, and eyes
appeared functionally developed. Total lengths were not
measured for every individual in this test but random
measurements indicated lengths of 12-14 mm for these groups.
A very strong preference for Side B as evidenced by initial
choices and distribution of total time was clearly demon-
strated by Groups A and C. Only Group B showed no prefer-
ence of first Side choice behavior. But this does not seem
a true representation of behavior as over 80 percent of the

total time was Spent in Side B.

 

26

H Him m6 m6 «.mm o .3 m.m.n .U

 

a n.@ m.oa m.mH m.sm H m H.ma .m

m m.HH m.s m.os m.m~ he s m.mm o

m o.HH m.m H.ms m.H~ ma m m.m~ m smusm >H

as H.om «.mm o.Hs s.mm as H s.mH 0

OH m.mm m.sm «.ms o.k~ es s «.ms m

m m.mv m.~m m.>m s.~m ms m o.ms a mmumm HH

m «.mm m.o~ m.ms m.s~ ms m o

o m.oo m.mm m.sm H.ma as s m

s o.mo o.os m.mm m.sa «a H m Haum H
mmmcmno m a m a m « Aesv.q.e muons Imcsnoumrumomv umme
seem memo

 

 

vamuw>fibom vamefle Hobos wowonu .uHcH

 

 

.mmaomomu mscmoanwem am no mcwummu onmocmum
cw momcmro stem one .>uw>fluom .weflu Hmuou .mooa0no asapflca mo cowusnauumflp sumo cmmz|u.H magma

27

Table 2.--Probability values of initial choice distributions
(chi-square test) and percent total time spent
over respective substrates (t-test) by B.
americanus larvae. ‘—

 

 

 

Test Group Initial Choices Total Time(%)

I A p < .001 p < .001
B p < .50 p < .001
C p < .01 p < .001

II A < .001 p < .001
B < .02 p < .001
C < .001 p < .001

IV B p < .005 p < .001
C p < .01 p < .01
B' p < .05 p < .001
C' p < .001 p < .001

 

28

Approximately two weeks subsequent to Test I, a
second test was run to determine if further conditioning or
development may have influenced choice behavior. No effect
was evidenced and all groups continued to strongly select
Substrate B in both categories of time and first choice.

As a consequence of a Sharp rise in laboratory
temperature soon after Test II was completed, tadpoles
developed very rapidly necessitating premature testing. At
the time of Test IV, all stages of metamorphic transforma-
tion were present, but due to losses from mortality and
metamorphosis an insufficient number of Group A tadpoles
were available for testing. Representative samples across
all stages were tested in an attempt to differentiate any
late developmental affects on habitat selection. The most
measureable characteristics during this period are the
degree of hind- and forelimb development and body-tail
proportions. During stages 43-44 (Gosner, 1960) when larvae
become frog-like and are nearing the juvenile, terrestrial
stage, a distinct reversal in habitat preference ensues.
Tadpoles with tail lengths measuring greater than 7 mm
continued to strongly select Substrate B. However, at tail
lengths equal to or less than 7 mm (Groups B' and C'),
previous behavior was reversed and Substrate A became the
preferred habitat.

It is apparent that rearing experience had no influ-
ence on initial choices in any test. To determine if there

was variation in distribution of the percent total times

29

among groups, one-way analyses of variance were calculated
and are presented for Test I, II, and IV in Tables 3, 4,
and 5 Showing sum of squares, mean squares, and F-values.
Results in all cases show total times independent of prior
habitat experience.

Activity over Substrates A and B in Tests I and II
was evaluated for dependent variables by subjecting data to
tests of two-way analyses of variance with replication.
Activity was designated as the percent of the total time
over each substrate Spent swimming. For Test I, Table 6
indicates effects arising from two sources. However, group
rearing experience, shown to be an influencing factor, is
misleading. A review of the data Showed variation pre-
dominately due to four individuals (three in Group A and
one in Group B). In these cases only three to 26 seconds
were spent in Habitat A of which the entire time was spent
active. This gave values of 100 percent activity over that
substrate, thus biasing values for that group. Greatest
Significance resulted from test substrate. Larvae of all
groups spent proportionately much less time active over the
avoided substrate (A). No interaction of experience and
test substrate was indicated. Table 7 illustrates similar
findings with activity data of Test II. Only test substrate
was responsible for variation. Because of minimal activity
in later stages, it was not possible to accurately evaluate

activity for larvae of Group IV.

30

Table 3.--One-way analysis of variance of percent total
time spent in Side A by E. americanus during

 

 

 

Test I.
Source of Variation df SS MS F
Among Groups 2 385.86 192.93 0.917 ns
Within Groups 42 8840.39 210.48
Total 44 9226.25

 

Table 4.--One-way analysis of variance of percent total
time Spent in Side A by B. americanus during

 

 

 

Test II.
Source of Variation df 88 MS F
Among Groups 2 296.78 148.39 0.624 nS
Within Groups 57 13,557.28 237.85

Total 59 13,854.06

 

31

Table 5.--One-way analysis of variance of percent total
time Spent in Side A for B, americanus tadpoles
with tail lengths greater than 7 mm and less than

or equal to 7 mm in Test IV.

 

 

Source of Variation df SS MS F
> 7 mm
Among Groups 1 490.86 490.86 0.784 ns
Within Groups 36 22,532.57 625.90
Total 37 23,023.43
.3 7 mm
Among Groups 1 374.04 374.04 1.687 ns
Within Groups 18 3,990.55 221.70
Total 19 4,364.59

 

Table 6.--Two-way anova with replication of the percent
total time spent active by B. americanus over

Habitats A and B in Test I.

 

 

 

Source df SS MS F
Subgroups 5 25744.72 5148.94
A (Experience) 2 5445.58 2722.79 4.170 p < .025
B (Habitat) 1 18454.18 18454.18 28.261 p < .001
A X B 2 1842.56 921.28 1.411 ns
Within (error) 84 54852.45 653.00
Total 89 80594.77

 

32

Table 7.--Two-way anova with replication of the percent
total time spent active by E. americanus over
Habitats A and B in Test II.

 

Source df SS MS F

 

Subgroups 5 14210.23 2842.05
A (Experience) 2 158.68 79.34 0.171 ns
B (Habitat) 1 11818.69 11818.69 25.421 p < .001
A x B 2 2232.86 1116.43 2.401 ns

Within (error) 114 53000.54 464.92

Total 119 67210.77

 

Results and behavioral observations of Tests I and
II suggested that the vertical stripes of Substrate A were
aversive, thereby affecting selection. In an effort to
resolve if there was actually positive orientation toward
Substrate B, Substrate A was replaced with a uniform white
substrate (C). Results and their significance are presented
in Table 8. Substrate B was clearly not preferred. On the
contrary, initial choices of Groups B and C and total time
distribution of Group B demonstrated preference for Sub-
strate C. Analysis of variance of total times (Table 9)

indicated no differences between groups.

Behavioral Observations
Each individual tested was observed and noted for
any stereotyped behavioral patterns in the test situation.

In Test I, avoidance behavior similar to that observed in

33

.mo. v m«««

 

 

.mwo. v m..
.moo. v as
m H.m~ o.sm m.ms H.Hm ..~ NH «.mm o
v s.mH s.ms ...m.sm ~.mo .H ms H.m~ m
as o.om m.mm m.ms H.sm m m spew a mmnmm
mmmcmno m o m o . m o Ieev.q.e mucus mama

 

 

 

mosm
sus>fluo¢ m mass dance a mosoro .uscH

 

 

.HHH umme SH 0 was m mumuflbmm coo3umb mnemowumam am an mmmcmno
mcwm ocm .>DH>Huom .mefiu Hmuou .mmowoso HMfluacw mo coflusnfluumflo muwo :m02||.m magma

34

Table 9.--One-way analysis of variance of percent total
time Spent in Side C by E. americanus during

 

 

 

Test III.
Source of Variation df SS MS F
Among Groups 2 1,405.96 702.98 1.741 ns
Within Groups 36 14,538.54 403.85
Total 38 15,944.50

 

R. gylvatica tadpoles was recorded, but in greater fre-
quency. Most of the time was spent along the extreme side
of Side B. Approaches to Side A and subsequent reversals
at the border were very common. Actual entries into Side A
most frequently resulted from approaches away from the sides
and toward the center of the tank. Less often, were pene-
trations resulting from approaches along one of the edges.
When this occurred, larvae would stereotypically redirect
movement upon entry at an approximate 45° angle away from
the Side bearing vertical stripes, cease movement, and come
to rest on the bottom. Tadpoles would usually return to
Side B within 30-60 seconds. Entries into Side A seldom
surpassed 3-6 cm. Activity within Side A, while proportion-
ately less than that in Side B, occurred away from the
sides of the tank and toward the center of that habitat.

In Test II, behavior varied little in frequency or
patterning from that observed in Test I. However, Group A

tadpoles exhibited much greater tendency to swim along the

35

Sides in Habitat A when tested. Active avoidance of the
vertical stripes was not as apparent in this group.

In Test III little stereotypic behavior was noted.
Larvae tended to select the uniform white substrate and
spent approximately 80-90 percent of their time in very
close proximity to the Sides of each habitat.

Most tadpoles tested in Test IV evidenced varying
stages of limb development and tail resorption. In these
stages of late development Spontaneous activity greatly
decreased in frequency and duration. AS a consequence,
"avoidance behavior" as described for Tests I and II was
uncommon.

Larvae equal to or less than 7 mm tail length
clearly reversed their previous avoidance of Side A.
Typically, initial movement would occur soon after release
and would consist of a direct, rapid movement toward the
distant side of Side A. At least 90 percent of all subse-
quent movement was composed of vertical swimming upward

against the sides of Habitat A and passive downward sinking.

Testing of Rana clamitans
Experimentation with Ragg_clamitans larvae proceded
in two directions. Testing for developmental and experi-
ential influences on choice of the standard substrates was
virtually unchanged from the procedures used with E.
americanus. A second phase of investigation was aimed at

the component of the artificial habitat perceived as

36

"aversive." This required increasing the density within
respective rearing tanks to 120 individuals, thus intensi-
fying the possible influence of density-dependent growth
inhibition. This added influence Should not significantly
alter the quality of desired data excepting possible time-
developmental interactions.
Testing of Conditioned Substrates:
Developmental and ExperientiaIi
Effects

Four tests using the two standard experimental sub-
strates were completed to elucidate developmental and
experiential affects. The general inactivity of these
larvae throughout development made it impractical to measure
total time Spent per Side and activity per Side. Therefore,
the sole indicator of habitat preference was initial choice
following release. The adequacy of this as a true indicator
was well established in previous testing of R. Sylvatica

and B. americanus.

 

Very early preference was tested at days 10-11
posthatching when Spontaneous swimming, feeding, and eye
development were sufficiently progressed. Subsequent tests
were conducted at 25, 52, 66, and 81 days posthatching.
Summary of the results and significance of initial choices
and prerelease orientation is given in Table 10.

At 10 and 11 days of age all groups exhibited very
strong preference of Side B. Excepting Group A, orientation

prior to release was also highly Significantly directed

37

 

 

 

 

m: m HH mo. ma v N.o¢ 0

ma ma b m0. ma v «.mv m

m: m va as n ma 0.0v d mmnam HHHx

m: h m Hoo. 0H 0 w.vm O mm x

m: o m Hoo. HH 0 m.mm O

we NH 0 0H. ma m m.mm m

mg m OH OH. ma m h.mm e vmnmm H>

m: m oa m: m 0H m.am O

we NH m we 0H m o.mm m

mo. v va 6 Ho. ma m «.mm a mmnmm HH

Ho. v Ha H Hoo. NH 0 v.ma 0

mo. v 0H m Ho. Ha H «.ma m

m: m m Ho. Ha H o.NH a HHIOH H
muHHAQMQoum m < >uaaflnmnoum m m AEEV.A.9 msouo Amcwnoumnumomv umme

mama
coflusucoauo mowonu .uacH
ommoaououm

 

moowono Howuflca mo cowusnwuumflo ocwsonn mcduwemao .

.mmfluwafinmnoum name» one mcofiumucoauo mmmmaoumum can
.m.mo muasmmu ummu pumocmum mo >HMEE:mtn.oa manna

 

38

toward Side B. Behavioral patterns within the test chamber
indicating avoidance of Side A were absent. At 25-28 days,
Group A followed Similar selection patterns in apparent
avoidance of its rearing substrate. However, Groups B and
C treated substrates nonpreferentially in both first choices
and prerelease positioning, indicating a developmental
change had occurred in preference in these groups.

Approximately four weeks later (Test VI), results
were less definitive. Group C strongly preferred Side B.
However, Groups A and B, though tending toward preference of
Side B, Showed no significant choice and none of the groups
indicated significant directional orientation prior to
release.

Following Test VI, there remained in conditioning
Habitats A and B only enough larvae for one more test.

These remaining larvae were subjected to an additional four
weeks experience within their rearing habitats. Approxim-
mately midway through this period, ten tadpoles of Group C
were tested to determine if any developmental change was
affecting habitat choice. No change was evidenced as all
tadpoles initially selected Side B.

On days 81-83, results show habitat conditioning
apparently affecting substrate choice in Group A. While
Group A larvae treated habitats nonpreferentially, Groups B
and C significantly selected Side B. The distribution of
choices for Group A is particularly noteworthy when compared

with choice behavior in the three previous tests.

39
Investigation of the
Aversive Compgnent

This series of tests were conducted with R.
Jolamitans larvae reared in an alternate control habitat. A
summary of the results is given in Table 11. Prerelease
orientations were randomly distributed in every case and
are not included in the summary.

Test III provided a choice between Habitats B and
C, as was carried out in testing of B. americanus tadpoles.
No significant selection behavior was demonstrated by any
group. However, it should be noted that in each group,
over twice as many larvae chose Side C over Side B, sug-
gesting initial choices may not be entirely randomly dis-
tributed.

Suggestions by these and prior experiments with R.
sylvatica and B. americanus were that the stripes of Side A
were in some way aversive. One of the most striking com-
ponents of this habitat to an observor is the sharp black-
white contrast. It was hypothesized that such a contrast
when positioned vertically could produce a highly aversive
"flicker" across the visual field of a swimming tadpole.
This hypothesis was simply tested by replacing the side
panels of Side A with horizontal stripes of the same
dimensions. Results Show that this procedure did not
eliminate the aversive quality of that habitat. Only

Group B did not strongly avoid this substrate.

40

.oosnuoo susmcmo.

 

:Hu¢\a mmnwm

 

 

 

 

 

moo. v es m o oomsuum .uuo> «.mo o omumm >x
=~\s.o\s moosm
soo. v ms m o oomsuum .uuu> m.ms 0 mo >sx
so. v ms s o o m.om o snuos ssx
m: ms 5 o Esmoosm .uuw> v.om u so sx
m: ms m o mooosm .uum> o.mm u so xs
mo os o o moosm gunsm o.~m o om sss>
me o m o moosm some m.mm 0 mm ss>
m: ss o «.mm o
m: ms o =suo\s moosw «.mm m
an o ss 0. oomsuum .uuw> o.~m s osuoo >
so. v ss s s.om 0
mo m m =~\s"~\s moosm o.om m
so. v ss s m oomsuum .usuom ~.sm a souos >s
m: m ss o.o~ o
m: m ss «.mm m
m: m ms m o s.m~ a mmusm sss
musssnmnoum ss s ss s 35 :s .,s. moons 85:33 some
-uaomo
mowono .uscH panama mumufinem when

 

was whososo Heavens mcs3osu mcmuwemao .

 

.mmsusssnsnoum usurp

m.>n :mmoso>smum>m= mswusmsuno>sw mums» mo mhmeesmnl.ss wanes

41

This subsequently left three major classes of com-
ponents of the vertical striping to investigate. The
pertinent stimulus could involve stripe width and/or
frequency, the overall light intensity of the substrate, or
the striped pattern itself. Test V retained basic pattern
and altered stripe widths to one-eighth inch black and one
inch white thus changing not only the ratio of black to
white, but also the overall reflective intensity of the
substrate. In view of the possible confounding influence
of the base stripes of Side B, in this and subsequent tests
novel substrates were tested against a uniform white sub-
strate. None of the groups demonstrated active differ-
entiation of these substrate types.

In consideration of these findings, it would be an
easy task to test if the change in overall substrate inten-
sity in Test V was responsible for these behaviors. To
accomplish this, sidepanels of Side A were removed exposing
the gray tank sides. This closely approximated the overall
intensity produced by the standard substrate. Group C tad-
poles behaved nonpreferentially in this Situation and when
presented with a habitat of uniform black replacing the
gray in Test VIII, likewise responded indifferently. This
indicated intensity alone was not a controlling influence
but could still be an important interaction with another
component.

In a study by Wiens (1970) with 5322 aurora tad-

poles, it was suggested that the striped habitat selected

42

by striped-conditioned tadpoles may have been analogous to
linear vegetation found in the natural habitat of this
Species. Langdon (1971) concluded that fiddler crabs, Egg
pugilator, approached vertical striping because of its
resemblance to protective grasses and other vegetation in
their habitat. On the same premise, a test was designed to
test the possible single influence of a natural occurring
linear plant on selection of a habitat by tadpoles. On two-
thirds of one side of an all white test tank, vertical
segments of Elodea canadensis were distributed evenly.
Individuals were released approximately one-third of the
way into the featureless side of the tank. In Test XI the
same basic procedure was followed but the density of Elodea
within Side I was roughly doubled. In neither case did the
presence of Elodea nor its density have any significant
influence on selection of a habitat.

A typical pond of southern Michigan differs little
in obvious visual features. One predominant variable factor
from an observor's standpoint, is the structural type and
density characteristics of aquatic vegetation. A simple
breakdown of vegetation type according to structure might
be those species of a pinnate or branched structure (like
Elodea) as opposed to plants arranged in linear uniformity.
In the present case, tadpoles avoid vertical, artificial
striping that is highly uniform, but do not react to actual
vegetation of a pinnate nature. In Test XII, Side A was

modified to test this basic difference in patterning. The

43

uniformity of type A substrate was disrupted by super-
imposing horizontal "fringe" on the black stripes (Figure 2).
This "fringe" component, however, was ineffective in
modifying avoidance behavior when tested against an all
white habitat. Only three of 18 larvae initially selected
Side D.

Further experimentation with patterns and new vari-
ables was impractical with the remaining number of larvae.
The best direction seemed to be with varying relative
amounts of black and white stripes without changing basic
pattern. In Test V, alteration of stripe widths and
reducing the black to white ratio to 1:8 resulted in no
preference. Because overall reflective intensity is not a
dominating influence, Test V results suggest tadpoles were
responding to width or frequency of either the black or
white stripes or to some relationship between the two
stripes in which a ratio is the important feature. In Test
XIV, it was decided to leave unchanged the width of the
black stripe (one-eighth inch) but to reduce the white
stripe width to that of the standard substrate (one-half
inch). If this habitat was treated nonpreferentially,
indications would be that either the black width or ratio
of black to white was important. If avoided, ratio may
still be important but in addition could mean the width of
the white stripe was the recognized feature. Results
showed avoidance was highly significant with only two of

20 Group C larvae selecting that habitat.

55858533883151?)

Figure 2. Test subst e used in Habitat D.

45

A final test (Test XV) was designed to elucidate
the importance of either ratio or white stripe width in
aversiveness. The avoided ratio of 1:4 was retained while
enlarging the white stripe to that employed in Test V (one-
fourth inch b1ack:one inch white). Definite avoidance was
again demonstrated, strongly suggesting that the ratio of
black to white is important in illiciting avoidance be-

havior in these larvae.

DISCUSSION

Develgpmental Affects

At two instances, results Showed habitat choice
behavior was strongly influenced by developmental stage. In
testing of Bpfg americanus tadpoles, selection of experi-
mental substrates was based on an inherent aversion to
Habitat A (striped Sides). Prior habitat experience had
no apparent influence. Preference for Habitat B by all
groups persisted throughout development until larvae prog-
ressed into late metamorphic transformation. On days 34-37
posthatching, those larvae of Groups 8' and C' with tails
resorbed to 7 mm or less reversed previous preference and
significantly selected Habitat A (Table 1).

At this stage, larvae have ceased feeding and
activity is greatly diminished from earlies stages. Tad-
poles appear frog-like with fore- and hind-limbs well
developed. Metamorphosis is progressing very rapidly with
almost complete tail resorption from 7 mm and conversion to
terrestrial life occurring within approximately 24 hours at
laboratory temperatures. There is also evidence of a
reversal in y-axis orientation occurring in tadpoles just

prior to emergence. McKeown (1968) found that larvae of

46

47

E, woodhousii and E. catesbeiana originally orient toward
deep water but Show Shoreward preferences prior to emer-
gence. A similar reversal just prior to emergence in bull-
frogs was recorded by Goodyear and Altig (1971). During
this period it is critical that they climb to a dry surface
and resume terrestrial feeding or drowning will result.

One would hypothesize that the mechanism for
accomplishing this end would require strong and inherent
attending to cues such as decreasing water depth and emer-
gent and submergent vegetation and possibly sun compass
orientation. Such a mechanism would primarily be based on
visual and tactile cues. (Indeed, Brockelman (1969) con-
sidered B. americanus larvae to be transformed when "they
showed a decided preference for solid substratum above the
water surface and when the tail was shorter than the hind
legs." This description would place tadpoles in the approx-
imate stage reversal of preference occurred. It is likely
that the dark mid-water stripes of Habitat A resemble a
"climbable surface" to the transforming larvae. Behavior
within the test chamber supports this contention.

Larvae with a tail length of 7 mm or less ex-
hibited stereotypic behavior following release. Upon
release, active choice was initiated usually within the
first one to two minutes. This almost invariably consisted
of direct swimming perpendicular from the border to the
extreme side of Habitat A. Once actual contact with the

side was made, subsequent swimming was nearly exclusively

48

in the vertical plane. Seldom was downward movement
actively directed, but rather consisted of passive sinking.
A second observation indicating a developmental

influence occurred with E. clamitans larvae of 25-28 days

 

of age at Test II (Table 10). A previous test at 10-11
days of age demonstrated strong preference of Habitat B by
all groups in both prerelease orientation and first choice.
At Test II, Group A continued to clearly prefer Habitat B
in both categories over its conditioning substrate. How-
ever, Groups B and C were randomly distributed in their
initial choices and prerelease orientations. Furthermore,
the change to nonpreference exhibited by Groups B and C was
transient in that both groups avoided Habitat A in later
tests. This does not seem to represent a procedurally
induced phenomenon as Group A tadpoles were tested at the
same time and subject to identical experimental and rearing
conditions. The functional or possible adaptive signifi-
cance of such a change in behavior is difficult to per-
ceive. There occurs no known change in morphology (except
size), feeding habits, or any other aspect of larval life
that may be critical to larvae and that might correspond
with the time period. Without more Specific information
about the larval ecology of this Species it is impossible
to interpret this lack of preference. Repetition of this

phase of the experiment would be highly desirable.

49

Experiential Affects

Rearing experience had no affect on habitat choice
in either 3. sylvatica tadpoles or during any period in the
larval life (34-37 days) of B, americanus. (Unfortunately,
the time to metamorphosis in the later group was Shortened
due to an increase in laboratory temperature.)

Two affects attributable to habitat conditioning
were both observed in standard testing of R. clamitans
larvae. On days 25-28 posthatching, green frog tadpoles of
Group A continued to avoid their rearing substrate while
Groups B and C showed no preference. Prior experience
within conditioning Habitat A by Group A was sufficient to
override a more general developmental change occurring in
Groups B and C. No explanation for this developmental
change is readily apparent from available data. It is also
difficult to determine if there is a true advantage to this
experiential influence. There seems to be no Situation
where benefit could be derived during this period of
development short of limited and transient spatial segre-
gation due to differential experiences.

Experientially modified behavior was also evidenced
late in the testing of B. clamitans Group A tadpoles.
Standard tests initiated at 10 and 25 days of age indicated
typical strong avoidance of its rearing substrate. At 52
days only 5 of 18 selected Habitat A (not significant at
p = .05). However, at days 81-83, 13 of 20 larvae tested

selected the conditioning substrate (A) while Groups B

50

and C Significantly avoided that substrate. This seems to
represent a period in which acceptance, or at least
acclimation, to the conditioned substrate is taking place.
This hardly seems an "adaptive" change occurring this late
in larval life and more probably is simply an habituation to
an inherently aversive stimulus.

Prior experience of B. americanus larvae had a
limited influence on behavior within the test chamber. In
Tests I and II movements by all groups illustrated a strong
avoidance to the vertical stripes of Side A. However, in
Test II, Group A larvae exhibited a much greater tendency
to swim along the Sides of Side A than the other groups that
were previously unexposed to that substrate. Exposure to
Habitat A within the first three weeks of larval life was
sufficient to decrease spontaneous reaction to the vertical
stripes but not sufficient to abolish avoidance of that

habitat.

Selection Based on Avoidance
Responses to standard substrates were relatively
stereotypic in choice and behavior in the three species
tested. Habitat B with the striped bottom was nearly
invariably preferred over the striped Sides of Habitat A.
Conditioning had little affect on modifying choice. The
primary motivator of this response seemed one of avoidance

to the vertical striping of Habitat A, rather than positive

51

orienting to Habitat B. Evidence for this is reflected
strongly in activity patterns within the test tank.

It was commonly observed in testing of R. sylvatica
and B. americanus that when Side A was approached from
Side B, tadpoles would rapidly reverse direction upon
reaching the border. If Side A was penetrated, distance of
penetration was very short, and movement, particularly by
E. americanus, was directed sharply away from the striped
Sides upon entry. Whereas time spent in Side B was almost
exclusively along the edges, time spent in Side A was toward
the center of that habitat away from the stripes. When
given a choice between Habitat B and an all white control
habitat (C), B. americanus and R. clamitans tended to prefer
Side C. This supports the contention that the vertical
stripes of Side A are the avoided component and that Side B
was in fact discriminated. Results also suggest that the
bottom stripes of Side B may be aversive but to a lesser
degree.

Spontaneous avoidance behavior of the vertical
stripes of Side A was not exhibited by larvae of R.

clamitans, but initial choices generally showed preference

 

for Side B. Evidence indicated that a similar selective
process (avoidance) was operating in this species as the
previous two, and several tests were designed to deduce the
Specific component(s) of Habitat A that was aversive

(Table 11) .

52

The possibility that aversive visual flicker was
responsible for avoidance of Habitat A was tested by
allowing tadpoles to choose between Habitat B and a habitat
bearing horizontal stripes in place of the vertical stripes.
No difference in behavior was noted as this substrate was
similarly avoided. This generally agrees with Munn'S
(1940) unsuccessful efforts at teaching H112 tadpoles to
discriminate between vertical and horizontal stripes.
Overall reflective intensity, as tested with uniform gray
and black sides against all white substrates, was in itself
not a factor influencing selection. Pattern disruption
facilitated by addition of a fringed component (Habitat D)
to the standard "A" substrate did not change its aversive
quality as evidenced by distribution of initial choices.

The only significant change in choice behavior
resulted from manipulation of relative proportions of verti-
cal black and-white stripes. Retention of basic pattern
and changing black stripe widths to one-eighth inch and
white stripe widths to one inch resulted in no preference
when tested against an all white substrate. Subsequent
tests of striping sought to resolve the components effective
in apparent "neutralization" of the aversive quality of
Habitat A. A second test doubled the nonpreferred stripe
ratio from 1:8 to 1:4 but retained the one-eighth inch
black stripe (one-eighth inch black, one-half inch white).
Avoidance was restored as Group C larvae strongly selected

the uniform white substrate.

53

These results suggested that either the single
effect of the white stripe width or the actual black-white
ratio was the feature affecting habitat selection. A final
test retaining the avoided 1:4 ratio while employing one
inch wide white intervals was designed to test these
factors. Avoidance of this substrate by Group C was highly
significant. Apparently the ratio of the contrasting black
to white'stripes is the aversive component affecting
observed selection in this species.

A number of problems arise in attempting to inter-
pret or assign adaptive function to the described responses.
The experimental visual parameters were highly artificial
in design. Ecological interactions and microdistribution
of anuran larvae are virtually uninvestigated. Sources of
mortality and the degree to which certain factors affect
development lack quantification. Furthermore, with the
exception of limited work with adults (Maturana et al.,
1960; Lettvin, et al., 1959; and Ewert, 1974) visual dis-
crimination and pattern resolution capabilities of anuran
larvae are unknown. In consequence, assumptions regarding
direct application of the described avoidance behavior to a
natural situation are purely speculative and have little
substantive basis.

The strength of the aversive stimulus, its per-
sistence in development, and its relative specificity in
three species of anurans demonstrate a strong hereditary

basis and suggest the behavior is being acted upon

54

selectively. For selection to be taking place, one must
assume a situation exemplifying similar qualities of dark-
light ratios currently exists within the natural distri-
bution of these animals and that avoidance of this feature
has a profound influence on some aspect of larval life
(e.g., in predatory or competitive interaction).

Direct predator resemblance is unlikely. The
striped pattern is not like any primary predator and tad-
poles show little reaction to visual "threats" around them
unless they are accompanied by actual mechanical distur-
bance. One of the few characteristics of southern Michigan
ponds approaching resemblance to the artificial components
is submergent vegetation type and/or density. Vegetation
type has been shown to be important in the selection of a
Stream habitat by sticklebacks (Baker, 1971; and Hagen,
1967). Wiens (1970) thought that the stripes Selected by
conditioned Raga aurora tadpoles may have resembled linear
vegetation of their natural habitat. Longdon (1971) drew
similar conclusions in his study of the fiddler crab, 223
pugilator. Avoidance on this basis to specific vegetation
type or density could be important to anuran larvae from a
competitive or predatory standpoint. However. the important
element of contrast seems to be lacking in the submerged
aquatic environment and there is no evidence indicating
spatial segregation correlating with vegetation in the
literature or from personal observation. Indeed, R.

clamitans larvae failed to actively discriminate two

 

55

densities of the common Elodea canadensis when tested
against uniform white substrates.

The only conceivable situation where visual cues of
high contrast and varying density could be perceived by
tadpoles is that of overhead emergent vegetation, or even
terrestrial vegetation, outlined against the sky. These
cues would in fact be within the visual field of these tad-
poles in respect to the dorsal positioning of the eyes and
could be instrumental in y-axis orientation.

An alternative application that warrants consider-
ation when working with amphibian larvae is the possibility
of relevance of these cues to adults. Unfortunately,
retention of learned or inherent behaviors beyond meta-
morphosis is virtually uninvestigated for amphibians. But
because of the drastic change of habitat from completely
aquatic to semi-terrestrial, it is possible that a function-
ing behavior of the type described could be present in the
larval stage without adversely affecting their survival and
come into use when the animal transforms into the terres-
trial stage. In juveniles and adults, overhead cues of the
type postulated above could become instrumental in short
range orientation toward bodies of water. Overhead
vegetation of high density might be avoided thus "directing"
movements toward more open areas where water might occur.
The more open areas would be unavoided or oriented toward.

Although such a mechanism alone would likely be of little

56

use, when used in conjunction with other recepted cues
(e.g., sun compass orientation) it could become important.

Ewert (1974) studied the neurophysiological basis
of orienting and avoidance responses to small objects in
adult Bpfg.ppfg, Behavioral response was dependent on the
object size, configuration, its angular velocity, and its
contrast with the background. There may be some relation-
ship between aspects of this visually guided behavior in
adults to pertinent visual stimuli in the larval stage.
Further investigation into this area is needed.

In view of the present knowledge concerning general
ecology of anuran tadpoles, there is no clear "functional"
explanation for the observed tadpole selection behavior.
The great difficulty lies in defining an environmental
parameter that reflects visually the recognized components
of the imposed experimental habitat. However, the observed
components may in fact be incomplete in representation of
the entire relevant stimulus or may even represent a super-
normal stimulus further confounding analysis. Perhaps the
best approach would be a neurophysiological breakdown of a
particular response (Lettvin, et al., 1959) and its direct
relationship to an innate behavior (Ewert, 1974). In any
case, if an adaptive function is sought, determination must
be made of the microdistributional relationships of anuran

larvae and the factors influencing survival.

57

Activity Over Test Substrates

AS an additional behavioral measure, the total time
spent active over respective test substrates was recorded
for R. sylvatica and B. americanus tadpoles. The extremely
low level of activity of 5. clamitans larvae prevented
accurate assessment of this relation with habitat in this
species.

The time an animal is active within particular
habitats has been interpreted as an indication of the
quality of that habitat as perceived by the animal. Sale
(1969b) based much of his theory of a ”selection mechanism"
on the level of activity or "appetitive exploration"
exhibited by an animal in one particular habitat-type.

Sale postulates a negative feedback system whereby under
adequate environmental conditions feedback intensity to the
"selection mechanism" will lower the appetitive exploration.
Magnitude of activity increases in more inadequate situ-
ations. He tested his theory using the surgeon fish,
Acanthurus triostegus, and recorded behavior corroborating
predictions based on the habitat selection mechanism.

Few researchers other than Sale have examined
activity as a component of habitat selection. However,
Harris (1952) and Wecker (1963) cite evidence that the
prairie deermouse, Peromyscus maniculatus is generally more
active in its preferred habitat. Quertermus (1972) found
no consistent correlation of activity with habitat prefer-

ence in his study of Tilapia mossambica.

58

In the present study, 3. sylvatica tadpoles Spent
approximately equal amounts of time active over each test
substrate. B. americanus, however, spent a great deal more
time active over the preferred substrate (Table 1).
According to Sale's model, one would predict greater
activity in nonpreferred Habitat A. However, other factors
warrant consideration. Activity within Habitat B was
concentrated almost entirely around the Sides of that
habitat. Also random observation of activity within rearing
tanks indicated movement to be primarily around the edges.
Within the test tank stripes appear to be avoided by the
tadpoles regardless of their positioning. Assuming this to
be true, the comparatively high activity levels while over
test Habitat B could result from the interaction of two
factors. The apparent tendency to move along edges of an
enclosure combined with an attempt to avoid the bottom
striping of Habitat B could be capable of increasing con-
siderably total activity over that substrate. In Habitat A,
tadpoles strongly avoid Side striping, thus displacing them
away from the edges of that habitat. Assuming interaction
with a vertical edge functions to "stimulate" activity,
total movement in this habitat would substantially lower.
If such an interaction is operating in this situation,
Sale's model has not been refuted.

It is also conceivable that following the typical

initial choice of Habitat B, subsequent entry into highly

59

aversive Habitat A is sufficiently startling or "Shocking"
to cause the observed immediate ceasing of movement. If
this is the case, direct comparison with Sale's results is
not possible. Sale's experiment measured only random move-
ments within a single habitat and would not account for
behavior within a highly aversive Situation which allows a

choice between two habitats. The movements of tadpoles in

the test bank could not be classified as random or "ex—

ploratory" in terms.of Sale's criteria.

Relative Im ortance of Midwater
andiBottom Stripes

Standard experimental substrates were designed to

 

 

test the relative importance of midwater and bottom cues in
selection of the test substrates. Because both habitats
utilized identical visual components, selection would
essentially reflect differences in positioning of stripes.
All species tested generally preferred Habitat B (bottom
stripes) to Habitat A (side stripes). However, the stronger
stimulus does not seem to be a positive attraction to
Habitat B, but rather avoidance of the vertical striping of
Habitat A as previously discussed.

Tadpoles discriminate each substrate and aversive-
neSS between the two substrates apparently differs only in
degree. When allowed to choose between an all white sub-
strate and Side B, B. americanus and R. clamitans larvae

tend to prefer the featureless substrate. Stimulus strength

60

and quality is likely affected in two ways. From the view-
point of a tadpole, the striping of Habitat B would present
less effective surface area impinging on the retina than
the vertical striping of Habitat A because of its bottom
positioning. Also because of positioning, stripes would
likely lack the distinctness of Habitat A and may appear
highly blurred to a tadpole sitting or moving along the

bottom.

A second consideration is the effective visual field
of the tadpoles. Eye placement in these species is dorsal.
Thus, the perceptual field of bottom cues is decidedly
decreased, due partly tothe lateral bulge of the body
ventral to the eyes. Eye positioning seems related to their
bottom dwelling habits. Heyer (1972) observed that of the
species of frogs studied in Thailand, midwater and surface
forms generally had laterally directed eyes while bottom
dwellers possessed dorsally directed eyes. This seems to
indicate the greater relative importance of midwater and
overhead cues to these species by virtue of their bottom
habits and effective visual field. In the present study,
midwater stripes provided the greatest stimulus in selection

of a habitat.

SUMMARY

Three species of anuran larvae were reared in
artificial habitats that varied only in visual
appearance. At different points in development,
samples of individuals from each habitat were
allowed to choose between two substrate types to
determine the influence of developmental stage and
prior experience on habitat selection. The design
of the standard test substrates allowed discernment
of the relative importance of midwater and bottom
cues in selection of a habitat. Results showed
highly stereotyped behavior dependent primarily on
cues situated in the vertical plane (midwater).

In a preliminary study, larvae of the wood frog,
Rgpa_sylvatica, were reared from the hatchling
stage in two experimental habitats. Habitat A was
composed of alternating vertical black and white
stripes of one-half inch widths affixed to the
Sides and a uniform white bottom. Habitat B con-
sisted of identical black and white striping on the

bottom and uniform white sides. Larvae were

61

62

presented with a choice between these two con-
ditioning habitats after 20-25 days experience.
Larvae conditioned in Habitat A (Group A) signifi-
cantly demonstrated preference for Habitat B in

both initial choice and total time. Of nine larvae
tested from Habitat B (Group B), no preferencekwas
evidenced by initial choice, however, total time
indicated a strong preference for Habitat B. During
testing, spontaneous reactions by individuals of
both groups when Habitat A was approached, indicated
that this selection was based on an avoidance to the
vertical stripes of that habitat.

Larvae of the American toad, BEES americanus, were
reared from hatchlings through metamorphosis in
experimental Habitats A and B and a uniform white
control habitat (Habitat C). When allowed to choose
between Habitats A and B on days 9-11 and 23-25
posthatching, all three groups demonstrated prefer-
ence for Habitat B in initial choice and total

time. Swimming behavior within the test apparatus
during each test indicated strong avoidance to the
vertical stripes of Habitat A. Activity for every
group was much greater while over preferred

Habitat B. This relationship of activity and

habitat is discussed in relation to other studies.

63

During the final standard test on days 34-37 post-
hatching, tadpoles were in varying stages of
transformation. Tadpoles of Groups B and C in early
metamorphosis with tail lengths greater than 7 mm
continued to strongly avoid Habitat A in initial
choice and total time. Those animals with tails
resorbed to 7 mm or less reversed choice behavior
and strongly selected Habitat A. This pronounced
developmental change occurs at a time when larvae
must soon leave the water if they are to survive
and seems to represent strong attending to cues.
resembling a climbable surface (vertical stripes).
This is substantiated by swimming behavior and
orientation within Habitat A during testing.

On days 32-33 larvae were presented with a choice
between Habitat B and control Habitat C. Groups
tended toward preference of Habitat C. This sub-
stantiated indications that the side striping of
Habitat A was the avoided component and suggested
that the bottom striping was not only discriminated,
but was also aversive in a lesser degree.

Larvae of the green frog, 5322 clamitans, were
reared from hatching in the three standard con-
ditioning habitats and tested for preference
through development. Because of very low levels of
Spontaneous activity, the sole indicator of habitat

selection was initial choice. On days 10-11, all

64

groups significantly chose Habitat B and also
demonstrated orientation toward that habitat
immediately prior to release. Group A tadpoles con-
tinued to demonstrate preference for Habitat B on
days 25-28 in both initial choice and prerelease
orientation. However, Groups B and C exhibited no
preference to the Standard substrates. It appears
that a general developmental change in preference
occurred at this stage and prior exposure of Group A
to Habitat A was sufficient to overcome this change.
There is no apparent explanation for this develop-
mentally influenced behavior. On days 52-54,

Groups A and B tended toward selection of Habitat B,
but not Significantly. Group C strongly preferred
Habitat B in this test and again on day 66. The
final standard test on days 81-83 posthatching,
showed Groups B and C avoiding Habitat A while
Group A indicated an influence on selection attri-
butable to prior conditioning by Showing no prefer-
ence.

Between standard tests with R. clamitans, a variety
of novel substrates, substituted in place of
Habitat A, were tested in an effort to determine
the "aversive" components of the vertical stripes.
Aversive visual flicker was eliminated as a factor
by testing with horizontal stripes. Overall

reflective intensity of the substrate was likewise

65

discarded as a Single influencer of selection by
testing uniform gray and uniform black Sides. Tad-
poles did not actively discriminate vertical seg-
ments of aquatic vegetation (Elodea canadensis)

and when horizontal "fringe" was added to the black
stripes of Habitat A, tadpoles continued to strongly
avoid that substrate. By manipulation of widths
and ratios of the vertical black and white stripes,
it was concluded that the ratio of the contrasting
stripes was important in eliciting avoidance to
Habitat A in this Species. Adaptive implications

of these behaviors are discussed.

LITERATURE CITED

LITERATURE CITED

Altig, R. and E. D. Brodie. 1972. Laboratory behavior of
Ascaphus truei tadpoles. J. Herpetol. 6(1):21-24.

Anderson, P. K. '1954. Studies in the ecology of the narrow-

mouthed toad, Microh la carolinensis carolinensis.
Tulane Stud. ZooI. 2525:15-46.

Baker, M. C. 1971. Habitat selection in fourspine stickle-
backs (Apeltes guadrasus). Amer. Midl. Nat. 85(1):
239-242.

Baldes, R. J. and R. E. Vincent. 1969. Physical parameters
of microhabitats.occupied by brown trout in an
experimental flume. Trans. Amer. Fish. Soc. 98(2):
230-238.

Bragg, A. N. 1940a. Observations on the ecology and
natural history of Anura I. Habits, habitat and
breeding of Bufo co natus Say. Amer. Nat. 74
(753,754):322-349 and 424-438.

Bragg, A. N. 1940b. Observations on the ecology and
natural history of Anura II. Habits, habitat, and
breeding of Bufo woodhousii woodhousii. Amer.
Midl. Nat. 24(2):306-32I.

Bragg, A. N. 1943. Observations on the ecology and
natural history of Anura XV. The hylids and
microhylids in Oklahoma. Great Basin Nat. 4:62-80.

Bragg, A. N. and C. C. Smith. 1942. Observations of the
ecology and natural history of Anura IX. Notes on
breeding behavior in Oklahoma. Great Basin Nat.
3:33-50.

Brattstrom, B. H. 1962. -Thermal control of aggregation
behavior in tadpoles. Herpetologica 18(1):38-46.

66

67

Brockelman, W. Y. 1969. An analysis of density effects
and predation in Bufo americanus tadpoles. Ecology
50(4):632-644.

Brown, L. E. and J. R. Pierce. 1965. Observations on the
breeding behavior of certain anuran amphibians.
Texas J. Sci. l7(3):313-317.

Butler, R. L. and V. M. Hawthorne. 1968. The reactions of
dominant trout to changes in overhead artificial
cover. Trans. Amer. Fish. Soc. 97(1):37-4l.

Ewert, J. P. 1974. The neural basis of visually guided
behavior. Sci. Amer. 230(3):34-42.

Goodyear, C. P. and R. Altig. 1971. Orientation of bull-
frogs (Rana catesbeiana) during metamorphosis.
Copeia I97I(2):362-364.

Gosner, K. L. 1960. A simplified table for staging anuran
embryos and larvae with notes on identification.
Herpetologica 16(3):183-190.

Hagen, D. W. 1967. Isolating mechanisms in threespine
sticklebacks (Gasterosteous). J. Fish. Res. Bd.
Canada 24(8):1 - .

Haines, T. A. and R. L. Butler. 1969. Responses of year-
ling smallmouth bass (Micropperus gplomieui) to
artificial shelter in a stream aquarium. J. Fish.
Res. Bd. Canada 26(1):21-31.

Harris, V. T. 1952. An experimental study of habitat
selection by prairie and forest races of the deer-
mouse, Peromyscus maniculatus. Contr. Lab. Vert.
Biol. Univ. Michigan 56:1-53.

Heatwole, H. 1961. Habitat selection and activity of the
wood frog, Rana S lvatica Le Conte. Amer. Midl.
Nat. 66(2):301-3I§.

Heatwole, H. 1962. Environmental factors influencing
local distribution and activity of the salamander,
Plethodon cinereus. Ecology 43(3):460-472.

Herreid, C. F. and S. Kinney. 1967. Temperature and
development of the wood frog, Rana gylvatica, in
Alaska. Ecology 48(4):579-590.

Heyer, W. R. 1973. Ecological interactions of frog larvae
at a seasonal tropical location in Thailand. J.
Herpetol. 7(4):337-361.

68

Hilden, O. 1965. Habitat selection in birds. Ann. 2001.
Fennici 2:53-75. .

Hunsaker, D., II and R. W. Crawford. 1964. Preferential
spawning behavior of the largemouth bass,
Micropterus salmoides. Copeia l964(l):240-24l.

Hutchison, V. H. 1958. The distribution and ecology of the
cave salamander, Eugycea lucifuga. Ecol. Monogr.
28(1):l-20.

Klopfer, P. H. 1963. Behavioral aspects of habitat
selection: the role of early experience. Wilson
Bull. 75(1):15-22.

KlOpfer, P. H. 1965. Behavioral aspects of habitat
selection: a preliminary report on stereotypy in
foliage preferences of birds. Wilson Bull. 77(4):
376-381.

Klopfer, P. H. 1969. Habitats and territories. Basic
Books, Inc., New York. 117 p.

Klopfer, P. H. and J. P. Hailman. 1965. Habitat selection
in birds. £3: Advances in the study of behavior.
Lehrman, D. S., R. A. Hinde, and E. Shaw [eds.].
Academic Press, N.Y. Vol. I, pp. 279-303.

Kwain, W. and H. R. McCrimmon. 1967. The behavior and
bottom color selection of the rainbow trout, Salmo
ggirdneri Richardson, exposed to different lig t
intensities. Anim. Behav. 15(1):75-78.

Lack, D. 1933. Habitat selection in birds with special
references to the effects of afforestation on the
Breckland avifauna. J. Anim. Ecol. 2:239-262.

Langdon, J. W. 1971. Shape discrimination and learning in
the fiddler crab Uca pugilator. Ph.D. Thesis,
Florida State University.

Lettvin, J. Y., H. R. Maturana, W. S. McCulloch, and W. H.
Pitts. 1959. What the frog's eye tells the frog's
brain. Proc. Inst. Radio. Engrs., N.Y. 47:1940-
1951.

Lucas, E. A. and W. A. Reynolds. 1967. Temperature
selection by amphibian larvae. Physiol. 2001.
40:(2):159-l7l.

Maturana, H. R., J. Y. Lettvin, W. S. McCulloch, and W. H.
Pitts. 1960. Anatomy and physiology of vision in
the frog (Rana pipiens). J. Gen. Physiol. 43(6):
129-176.

69

McCrimmon, H. and W. Kwain. 1966. Use of overhead cover
by rainbow trout exposed to a series of light in-
tensities. J. Fish. Res. Bd. Canada 23(7):983-990.

McKenzie, D. S. and R. M. Storm. 1970. Patterns of habitat
selection in the clouded salamander, Aneides
ferreus (Cope). Herpetologica 26(4):450-454.

McKeown, J. P. 1968. The ontogenetic development of y-
axis orientation in four species of anurans. Ph.D.
Thesis, Mississippi State University.

Mullally, D. P. 1953. Observations on the ecology of the
toad Bufo canorus. Copeia l953(3):182-183.

Munn, N. L. 1940. Learning experiments with larval frogs.
J. Comp. Physiol. Psychol. 29:97-108.

Pache, J. 1932. Formensehen bei Froschen. Zeit. Vergl.
Physiol. 17:423-463.

Quertermus, C. J. 1972. Prior experience as a factor in
habitat selection of the cichlid fish Tila ia
mossambica. Ph.D. Thesis, Michigan State University.

Ralin, D. B. 1968. Ecological and reproductive differ-
entiation in the cryptic species of the H la
versicolor complex (Hylidae). Southwest. at.
I3(3):283-300.

Sale, P. F. 1968. Influence of cover availability on
depth preference of the juvenile manini, Acanthurus
triostegus sandvicensis. Copeia l968(4):802-807.

Sale, P. F. 1969a. Pertinent stimuli for habitat
selection by the juvenile manini, Acanthurus
triostegus sandvicensis. Ecology 50(4):616-623.

Sale, P. F. 1969b. A suggested mechanism for habitat
selection by the juvenile manini Acanthurus
trioste us sandvicensis Streets. _Behaviour 35:
27—44.

Sale, P. F. 1971. Apparent effect of prior experience on
a habitat preference exhibited by the reef fish,
Das llus aruanus (Pisces: Pomacentridae). Anim.
e av. 9(2):251—256.

Sale, P. F. 1972. Influence of corals in the dispersion
of the pomacentrid fish, Dascyllus aruanus.
Ecology 53(4):74l-744.

70

Sargent, T. D. 1965. The role of early experience in the
nest building of the zebra finch. Auk 82(1):48-61.

Savage, R. M. 1952. Ecological, physiological and
anatomical observations on some species of anuran
tadpoles. Proc. 2001. Soc. Lond. 122:467-514.

Sexton, O. and E. P. Ortleb. 1966. Some cues used by the
leptodactylid frog, Engystomops pgstulosus, in
selection of the oviposition site. Copeia 1966(2):
225-230.

Sokal, R. R. and F. J. Rohlf. 1969. Biometry. The
principles and practice of statistics in biological
research. W. H. Freeman and Co., San Francisco.

Svardson, G. 1949. Competition and habitat selection in
birds. Oikos 1:157-174.

Thoms, C. L. 1974. Effects of prior experience upon
nesting and resting site selection by the Mossambique
mouth brooder, Tila ia mossambica. Ph.D. Thesis,
Michigan State University.

 

Vlaming, V. L. de and R. B. Bury. 1970. Thermal selection
in tadpoles of the tailed-frog, Ascaphus truei. J.
Herpetol. 4(3-4):179-189.

Wecker, S. C. 1963. The role of early experience in
habitat selection by the prairie deer mouse,
Peromyscus maniculatus bairdi. Ecol. Monogr. 33(4):

Wells, M. M. 1924. Chrophilus nigritus. The swamp tree

frog. Turtox News :

Wiens, J. A. 1970. Effects of early experience on sub-
strate pattern selection in Rana aurora tadpoles.
Copeia l970(3):543-548.

Wiens, J. A. 1972. Anuran habitat selection: early
experience and substrate selection in Rana cascadae
tadpoles. Anim. Behav. 20(2):218-220.

Wilbur, H. M. 1972. Competition, predation, and the
structure of the Ambystoma-Rana gylvatica community.
Ecology 53(1):3-21T

71

Wilbur, H. W. and J. P. Collins. 1973. Ecological aspects
of amphibian metamorphosis. Science 182(4119):
1305-1314.

Workman, G. and K. C. Fischer. 1941. Temperature selection
and the effect of temperature on movement in frog
tadpoles. Amer. J. Physiol. 133:P499-P500.

HICHIGQN STRTE UNIV. LIBRRRIE

lll4|llllllLlllllllllllflllllllllllllHllillllllllllllllllllllll

2930069701 19