Witherington and Bjorndal 1991a Influences of wavelength and intensity on hatchlingsInfluences of Wavelength and Intensity on Hatchling Sea Turtle Phototaxis: Implications for
Sea-Finding Behavior
Author(s): Blair E. Witherington and Karen A. Bjorndal
Source: Copeia, Vol. 1991, No. 4 (Dec. 13, 1991), pp. 1060-1069
Published by: American Society of Ichthyologists and Herpetologists (ASIH)
Stable URL: http://www.jstor.org/stable/1446101 .
Accessed: 29/05/2014 13:28
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
.
American Society of Ichthyologists and Herpetologists (ASIH) is collaborating with JSTOR to digitize,
preserve and extend access to Copeia.
http://www.jstor.org
This content downloaded from 164.51.49.160 on Thu, 29 May 2014 13:28:38 PM
All use subject to JSTOR Terms and Conditions
Copeia, 1991(4), pp. 1060-1069
Influences of Wavelength and Intensity on Hatchling Sea Turtle
Phototaxis: Implications for Sea-Finding Behavior
BLAIR E. WITHERINGTON AND KAREN A. BJORNDAL
Visual cues are important to sea turtle hatchlings in determining seaward
direction upon emerging from the nest. In this study, we examined the roles that
color and intensity play in the sea-finding mechanisms employed by loggerhead
(Caretta caretta) and green turtle (Chelonia mydas) hatchlings. We tested hatch-
ling preference for a standard source of constant intensity and color (1.26 x
101" photons s-' m-2 at 520 nm), versus an adjustable light source (one of five
monochromatic colors at each of seven photon intensities), using a two-choice
apparatus. Both species oriented toward near-ultraviolet (360 nm), violet (400
nm), and blue-green (500 nm) light but chose the standard light source over
yellow-orange (600 nm) and red (700 nm) light. There was a positive relationship
between intensity and preference with 360, 400, and 500 nm light. We also
examined hatchling choice of either a darkened window or a window lighted by
one of eight monochromatic colors at each of two intensities. In these experi-
ments, loggerheads oriented toward 360, 400, and 500 nm light but away from
light in the green-yellow to yellow-orange range (560, 580, and 600 nm). Log-
gerheads oriented toward 700 nm light only at high intensity. Green turtles
responded insignificantly to 600 or 700 nm light at either intensity. The contrast
of green turtle behavioral responses with published electrophysiological data
and the aversion to yellow light observed in loggerheads suggest some level of
spectral quality assessment in sea finding for both species.
H ATCHLING sea turtles emerge from sub-
surface nests on oceanic beaches, pri-
marily at night, and immediately move toward
the sea. Hatchlings not entering the ocean ex-
peditiously suffer high mortality from preda-
tion, exhaustion, and desiccation. The sea-find-
ing behavior of neonate hatchlings occurs to
the exclusion of other predator avoidance be-
haviors. The robust nature of the sea-finding
response in sea turtles makes it an excellent
model for the study of animal orientation. A
detailed description of hatchling sea-finding be-
havior in the green turtle (Chelonia mydas) is
given by Carr and Ogren (1960).
Bilaterally blindfolded green turtle (Carr and
Ogren, 1960) and loggerhead (Caretta caretta;
Daniel and Smith, 1947) hatchlings are unable
to orient directly toward the sea, providing the
best evidence that hatchling sea finding is de-
pendent on visual cues. Green turtle hatchlings
also are attracted to artificial light sources (Mro-
sovsky and Shettleworth, 1968) and lightly tint-
ed objects (Carr and Ogren, 1960) and will move
in those directions irrespective of beach slope.
The use of vision by sea-finding hatchlings may
include an assessment of any combination of the
many properties of light detectable by the eye.
Properties of light studied with respect to hatch-
ling orientation include intensity, color, direc-
tion, and shape.
Two models predict that sea-finding hatch-
lings will move in the brightest direction.
Brightness in this sense is a measure of intensity
having both the directional and spectral sensi-
tivity properties of the hatchling in question.
Models describing the mechanism by which
hatchlings orient in the brightest direction in-
volve either phototropotaxis (Mrosovsky and
Kingsmill, 1985) or a direction system (Verhei-
jen and Wildschut, 1973; van Rhijn, 1979). A
third model involves shape as a visual cue (Par-
ker, 1922; Limpus, 1971; van Rhijn, 1979). This
model implies form vision and the ability to
recognize differences of pattern in the silhou-
ettes of duneward and seaward horizons. Ex-
periments conducted by van Rhijn and van Gor-
kom (1983) suggest this system may be distinct
from, but complementary to, either of the two
previous systems.
Spectral quality assessment, or the use of col-
or as a seaward orientation cue also has been
suggested (Hooker, 1911; Mrosovsky, 1972).
Differential responses to colors may be because
of color discrimination as well as a spectral bias
? 1991 by the American Society of Ichthyologists and Herpetologists
This content downloaded from 164.51.49.160 on Thu, 29 May 2014 13:28:38 PM
All use subject to JSTOR Terms and Conditions
WITHERINGTON AND BJORNDAL-SEA TURTLE PHOTOTAXIS 1061
resulting from physiological spectral sensitivity
range. Both light intensity and spectral prop-
erties play important roles in each of the visual
sea-finding mechanisms above. Previous at-
tempts to elucidate the roles of color and in-
tensity in sea turtle hatchling orientation have
been limited (Mrosovsky and Carr, 1967; Mro-
sovsky and Shettleworth, 1968). Primary limi-
tations have been an inability to isolate narrow
bandwidths of spectral light and to measure light
intensity (e.g., irradiance) at specific wave-
lengths. Broad-band filters used in earlier stud-
ies allow the transmission of many wavelengths
and make the assignment of hatchling response
to specific colors questionable.
In this paper, we assess preferences for light
intensities and narrow bandwidth color sources
in loggerhead and green turtle hatchlings dur-
ing their sea-finding behavior. We evaluate the
spectral quality assessment model for hatchling
sea finding and suggest a reassessment of mod-
els that predict "brightest direction" orienta-
tion.
METHODS
Hatchlings. -Loggerhead and green turtle
hatchlings for the experiments were taken from
clutches transferred into a secured hatchery near
Melbourne Beach, Florida, during the summer
nesting season, 1988. We examined hatchery
nests for signs of hatchling emergence activity
at dusk beginning 50 d into incubation (incu-
bation period approximately 50-57 d). Hatch-
lings typically lie just beneath the surface of the
sand until decreasing nighttime temperatures
prompt their emergence en masse (Mrosovsky,
1968). We collected hatchlings just after dusk,
when they were ready to emerge, and trans-
ported them in darkened buckets to an indoor
laboratory within 200 m of the hatchery. We
kept hatchlings in the dark to ensure that they
remained dark adapted and photically naive for
experimental trials. Hatchlings were collected
and used in experiments during a time when
hatchlings are normally emerging and moving
to the sea (2100-0200 h, Witherington et al.,
1990). Each hatchling was used for a single trial
and released the same night on the beach.
Experimental apparatus.-We used a modified
T-maze (V-maze) to determine hatchling pref-
erence with respect to photic cues (Fig. 1). This
apparatus was a V-shaped wooden box, with
each identical arm 78 cm in length. Openings
"L
-IF
"NDF
SL
Fig. 1. Schematic diagram of a V-maze used in
two-choice, color preference experiments with log-
gerhead and green turtle hatchlings. L = lamp; I =
iris; IF = interference filter; NDF = neutral density
filter; W = window; P = pitfall; SL = standard lamp.
at either end of the V-maze were 32 x 32 cm
and covered with windows of diffusing acrylic.
Black, flocked paper lined the inside of the box.
Hatchlings were introduced into the V-maze
through a 32 x 32 cm opening in the top of
the box near the vertex, which was covered with
a black cloth curtain. Hatchlings were covered
with an opaque cup that was raised after ap-
proximately 20 s to release the hatchling for
each trial. We placed hatchlings within the cup
so that they pointed toward the midpoint be-
tween the two windows, although hatchlings of-
ten altered their orientation during the time
preceding release. Hatchlings walking from the
vertex down either alley of the V-maze would
fall into a cloth pitfall pocket at the base of
either window.
We positioned two light sources so they shone
through each of the two windows of the V-maze
(Fig. 1). One of the light sources-hereafter
termed standard source-emitted light of con-
stant color (peak emission 520 nm) and intensity
This content downloaded from 164.51.49.160 on Thu, 29 May 2014 13:28:38 PM
All use subject to JSTOR Terms and Conditions
1062 COPEIA, 1991, NO. 4
(1.26 x 10'1 photons s-' m-2 at 520 nm). This
source was a conventional tungsten lamp pow-
ered with a 3.0 V DC source and fitted with a
blue gel filter. We used the blue filter to reduce
intensity of the standard source and to make
the color of the source more easily reproduced.
This standard source attracted all hatchlings (n
= 20) of each species to a lighted V-maze win-
dow when no other light was presented in pre-
trials. A second source-hereafter termed ad-
justable source-could be varied in color and
intensity. Light for this source originated from
a 3200 K (manufacturer's specification) tung-
sten lamp operated at 115 V AC. We regulated
spectral emission of the source with narrow band
interference filters (Melles Griot, half band-
width = 10 nm; transmission outside 20 nm
bandpass <0.001%). The angle of the light
source beam with the filters was 90* for all trials.
We regulated intensity by use of an iris aperture
and combinations of neutral density filters
(Melles Griot, 3.0, 2.0, 1.0, 0.5, 0.3, 0.1, and
0.04 OD). Intensity was measured as photon
flux at the hatchling release point with a LICOR
LI-1800 spectroradiometer.
We conducted experiments in a room com-
pletely darkened except for the standard and
adjustable sources. A series of light baffles pre-
vented light of either source from affecting the
other window. Line voltage for the sources was
monitored and did not vary more than 1% dur-
ing experiments.
Treatments.-To determine whether hatchling
orientation might be biased toward one of the
two windows, we released 42 hatchling logger-
heads and 23 hatchling green turtles within the
V-maze when both light sources were off. The
distribution of loggerhead and green turtle
hatchlings falling into pitfalls at either window
of the V-maze with both light sources off could
not be distinguished from random (alpha = 0.05,
binomial probability test, Z = 0.179 and Z =
0.133, respectively).
To determine the relative preference of
hatchlings for light of specific color and inten-
sity, we released hatchlings within the V-maze
while respective windows were illuminated with
the adjustable source and standard source. In
these trials-identified as adjustable source ver-
sus standard source-the adjustable source var-
ied among five monochromatic colors and six
intensities. Treatment colors were 360 nm (near-
ultraviolet), 400 nm (violet), 500 nm (blue-
green), 600 nm (yellow-orange), and 700 nm
(deep red) peak transmission. We measured
treatment intensities as photon flux and as-
signed values to a logarithmic scale. Intensities
were (log values shown parenthetically) 2.50 x
10'4 (0.7), 1.27 x 10' (1.4), 6.31 x 10' (2.1),
3.17 x 10'16(2.8), 1.58 x 10'7 (3.5), and 1.44
x 10'9 (5.5) photons s-I m-2 at respective peak
wavelengths. Log intensity 0.7 at 500 nm ap-
proximated the illuminance level measured for
a moonlit night. Higher light levels are com-
parable to those at dawn or dusk, times at which
hatchlings also must locate the sea. Because of
the emission spectrum of the incandescent
source, the highest log intensity (5.5) could only
be reached at the longest wavelengths (600 and
700 nm). We ran one set of trials for each spe-
cies with the adjustable source off (intensity =
0.0). Combinations of wavelengths and inten-
sities constituted 28 experimental treatments.
Thirty loggerhead hatchlings, each from a dif-
ferent clutch, were used individually for each
treatment. Ten green turtle hatchlings were
used in each treatment and originated from
three separate clutches.
To determine the polarity of hatchling re-
sponse (attraction or avoidance) to light of spe-
cific color and intensity, we released hatchlings
within the V-maze with one window illuminated
by the adjustable source only. In these trials-
identified as adjustable source versus darkened
window-the standard source window re-
mained dark. We used eight monochromatic
colors for the adjustable source in these treat-
ments: 360, 400, 500, 540 (yellow-green), 560
(green-yellow), 580 (yellow), 600, and 700 nm
peak transmission. Adjustable source intensity
for one set of eight experimental treatments
was 3.5 on the log scale. In a second set of eight
treatments, we used maximum source intensity,
and log intensity of the source varied according
to the maximum emission of the incandescent
lamp at each wavelength. Maximum intensity
of the source at each wavelength was 3.5 (360
nm), 4.0 (400 nm), 4.8 (500 nm), 5.0 (540 nm),
5.2 (560 nm), 5.3 (560 nm), 5.5 (580 nm), 5.5
(600 nm), and 5.7 (700 nm) log units. We ran
additional treatments with the adjustable source
off for 17 treatments total. Thirty loggerhead
hatchlings, three from each of 10 clutches, were
run individually per treatment. We used 20
green turtle hatchlings from a single clutch for
only the 600 and 700 nm treatments at maxi-
mum intensity.
In all trials, we excluded from the analysis
hatchlings not falling into either pitfall within
This content downloaded from 164.51.49.160 on Thu, 29 May 2014 13:28:38 PM
All use subject to JSTOR Terms and Conditions
WITHERINGTON AND BJORNDAL-SEA TURTLE PHOTOTAXIS 1063
2 min following their release. Less than 10% of
the 1523 loggerhead and 384 green turtle
hatchlings were excluded. In one to five trials
conducted for each treatment, we observed the
behavior of hatchlings within the V-maze
through the opening at the top of the box. In
treatments with both sources darkened, hatch-
lings were observed through active system night
vision goggles using an infrared source mount-
ed above the opening. Data from these trials
were not included in the analysis of source pref-
erence. Statistical tests for nominal data were
used with a null hypothesis rejection criterion
of alpha = 0.05.
RESULTS
Hatchling behavior.-In trials resulting in a
hatchling choosing one of the two windows of
the V-maze, hatchlings exhibited a head-up pos-
ture prior to and during movement. Whereas
loggerhead hatchlings typically paused prior to
moving, green turtle hatchlings were more ac-
tive, in some cases appearing to move in what-
ever direction they initially faced. Loggerhead
hatchlings moving toward lighted sources typ-
ically made direct movements without circling
(27 of 29). In trials with both sources off, hatch-
lings often circled (6 of 10) and made their way
to either pitfall by walking along the walls of
the box.
Loggerhead hatchlings that fell into the dark-
ened window pitfall in trials with 600 nm light
at the adjustable source window did not travel
directly to the darkened window. In all of 35
trials observed, hatchlings appeared to turn away
from the 600 nm lighted window and walk
against the opposite wall, attempting to climb,
until they fell into the dark window pitfall. In
30 additional trials, we placed loggerhead
hatchlings halfway from the maze vertex to the
600 nm source window (log intensity 3.5) facing
the 600 nm source. All of these hatchlings turned
away from the lighted window and moved di-
rectly to the opposite wall.
Adjustable source versus standard source.-We
found a positive relationship between log inten-
sity and the number of loggerhead hatchlings
preferring the adjustable source in the shorter
wavelength treatments: 360, 400, and 500 nm
(Fig. 2). This relationship did not differ among
these three treatments (chi-square = 1.86, df =
8). At longer wavelengths, 600 and 700 nm, the
number of hatchlings choosing the adjustable
source was not statistically different from zero
(binomial probability test) at all intensity levels.
With the adjustable source at log intensity 0, all
loggerhead hatchlings chose the standard
source.
The behavioral response of green turtle
hatchlings to spectral light (Fig. 3) was similar
to that of loggerhead hatchlings (Fig. 2). Sample
sizes do not allow chi-square comparisons be-
tween green turtle and loggerhead distribu-
tions. Green turtle hatchlings showed little pref-
erence for 600 and 700 nm light at any intensity
and increasing preference for 360, 400, and
500 nm light with increasing log intensity (Fig.
3).
Adjustable source versus darkened window.-Log-
gerhead hatchlings chose the window lighted
with 360 nm (binomial probability test, Z = 6.25),
400 nm (Z = 6.25), or 500 nm (Z = 5.43) light
at log intensity 3.5 significantly more often than
the darkened window (Fig. 4a). Conversely, log-
gerheads presented 560 nm (Z = -5.07), 580
nm (Z = -5.43), or 600 nm (Z = -6.25) light
at 3.5 intensity versus a dark window predom-
inantly chose the dark window (binomial prob-
ability test). In the 540 nm (Z = - 1.02) and 700
nm (Z = 1.30) trials, loggerhead hatchling pref-
erence for the adjustable source window or dark
window could not be distinguished from ran-
dom (binomial probability test). Loggerhead
preference for the two windows, with the ad-
justable source at maximum intensity, was sig-
nificantly different from random in all treat-
ments (360 nm, Z = 6.25; 400 nm, Z = 6.25;
500 nm, Z = 5.43; 540 nm, Z = 3.01; 700 nm,
Z = 4.56) except for the 560 nm (Z = 0.08) and
580 nm (Z = 1.35) trials (Fig. 4b, binomial prob-
ability test). In trials with 360, 400, 500, 540,
or 700 nm light, the lighted window was sig-
nificantly preferred, but in 600 nm trials (Z =
-6.25), all loggerhead hatchlings chose the dark
window over the lighted window (binomial
probability test).
In the only trials of this type run with green
turtles, more hatchlings chose the lighted win-
dow when presented 600 nm (n = 14 of 20) and
700 nm (n = 11 of 20) light at maximum inten-
sity, but the distributions cannot be distin-
guished from random (Z = 1.52 and Z = 0.471,
binomial probability test). Loggerhead and
green turtle preference for either window of
the V-maze in trials with both windows darkened
was not different from random (loggerheads, n
= 30, Z = 0.520; green turtles, n = 30, Z =
0.252; binomial probability test).
This content downloaded from 164.51.49.160 on Thu, 29 May 2014 13:28:38 PM
All use subject to JSTOR Terms and Conditions
1064 COPEIA, 1991, NO. 4
WAVELENGTH
30
- 360 NM
25 400 NM
500 NM
- 20 --0-- 600 NM
•
-- 700 NM
15
LL
O
e 10
D 5
0
0 1 2 3 4 5 6
LOG RELATIVE PHOTON FLUX
Fig. 2. The number of loggerhead hatchlings out of 30 choosing a light source varying among the
wavelengths and intensities specified, over a source of constant color and intensity. All represented hatchlings
chose one of the two sources. Colors are near-ultraviolet (360 nm), violet (400 nm), blue-green (500 nm),
green (540 nm), green-yellow (560 nm), yellow (580 nm), yellow-orange (600 nm), and deep red (700 nm).
DiscussION
Range of spectral sensitivity.-This study gives an
indication of a minimal spectral sensitivity range
for loggerheads and green turtles. Detailed in-
formation on the spectral sensitivity of sea tur-
tles exists only for the green turtle (Ehrenfeld,
1968; Granda and Dvorak, 1977). The action
spectrum based on electrophysiological data
provided by Granda and O'Shea (1972) shows
a greater spectral sensitivity in the shorter wave-
lengths but does not extend to wavelengths
shorter than 400 nm. Work by Ehrenfeld (1968)
with green turtles demonstrates that adult fe-
males can locate the sea after nesting when fit-
ted with goggles transmitting primarily near-
ultraviolet light (300-400 nm). The goggles in
these experiments did, however, transmit ap-
proximately 1% of the light in the 400 to 700
nm range, a substantial amount for the daytime
trials conducted. Because of the narrow spectral
bandwidths transmitted by the interference fil-
ters we used and the minimal leakage outside
the specified spectral bands, we can assign min-
imal spectral limits ? 10 nm to green turtle and
loggerhead vision. We have shown that green
turtle vision extends at least from the near-ul-
traviolet (360 nm) to green (500 nm). Vision in
loggerhead turtles extends minimally from the
near-ultraviolet (360 nm) to the red (700 nm).
One cannot quantify physiological spectral
sensitivities using our preference data because
of possible behavioral bias in the way the turtles
react to spectral light. A naive analysis of the
first series of treatments (Fig. 2), for instance,
may discount the ability of loggerheads to see
longer wavelengths (600 and 700 nm), an ability
demonstrated by the nonrandom choice of the
dark window in trials with 600 nm light (Fig.
4a-b) and the lighted window in trials with 700
nm light (Fig. 4b). The nonrandom response of
loggerhead hatchlings to 700 nm light only at
log intensity 5.7 may indicate that this wave-
length borders the spectral sensitivity range of
this animal. It is also possible that hatchlings
responded to light of other wavelengths trans-
This content downloaded from 164.51.49.160 on Thu, 29 May 2014 13:28:38 PM
All use subject to JSTOR Terms and Conditions
WITHERINGTON AND BJORNDAL-SEA TURTLE PHOTOTAXIS 1065
WAVELENGTH
10
--- 360 NM
8 - 400 NM
gl.
* 500 NM
--- 600 NM
--- 700 NM
I,-,, LL
O 4-
LI4
2
0 1 2 3 4 5 6
LOG RELATIVE PHOTON FLUX
Fig. 3. The number of green turtle hatchlings out of 30 choosing a light source varying among the
wavelengths and intensities specified, over a source of constant color and intensity. All represented hatchlings
chose one of the two sources.
mitted by the 700 nm filter, although this light
would have been more than four orders of mag-
nitude less intense than the 700 nm light.
In comparison with the pond slider Trachemys
scripta, vision in sea turtles extends farther into
the shorter wavelengths (near-ultraviolet) and,
at least in the case of the green turtle, dimin-
ishes more abruptly in the longer wavelengths
(Granda and Dvorak, 1977). Heightened sen-
sitivity in the near-ultraviolet and other short
wavelengths might be characterized as an ad-
aptation for vision in seawater where longer
wavelength light attenuates more abruptly with
depth (Loew and Lythgoe, 1985).
Spectral light and sea-finding.-A comparison of
the green turtle action response measured by
Granda and O'Shea (1972) with our measure-
ments of behavioral response to spectral light
(Fig. 5) reveals a behavioral bias against longer
wavelength light. Care should be taken in draw-
ing conclusions from the comparison in Figure
5. The electroretinogram (ERG) and behavioral
data were taken in different ways, although each
constitutes a spectral sensitivity of sorts. The
ERG curves are adapted from work by Granda
and O'Shea (1972) and illustrate light intensi-
ties at each color necessary to evoke respective
high and low criterion voltage responses at the
dark-adapted green turtle eye. The behavioral
curve is from the present study and illustrates
the number of dark-adapted green turtle hatch-
lings choosing each color at a single intensity
over a light source of standard intensity and
color. The behavioral curve is from the data of
Figure 3 for log intensity 3.5. The trend shown
in Figure 5 is similar for all light intensities in
the green turtle behavioral experiments we con-
ducted (Fig. 3). That is, whereas both ERG
curves show a peak sensitivity surrounding 600
nm, light in the 600 nm range is relatively un-
attractive to orienting green turtle hatchlings
This content downloaded from 164.51.49.160 on Thu, 29 May 2014 13:28:38 PM
All use subject to JSTOR Terms and Conditions
1066 COPEIA, 1991, NO. 4
CONTROL
700
* 600
* 580
* 560 a
540
* 500
Z * 400
I-z
Zj 360
z CONTROL
- J
S * 700
* 600
580
560 b
* 540
* 500
* 400
* 360
0 3 6 9 12 15 18 21 24 27 30
HATCHLINGS CHOOSING ADJUSTABLE SOURCE
Fig. 4. The number of loggerhead hatchlings choosing an adjustable light source varying among the
wavelengths specified, at an intensity of 3.5 log units (a) and at maximum source intensity (b). A darkened
window served as the alternate choice. During the control treatment, lighting remained off. All represented
hatchlings chose either the darkened window or the light source. Distributions marked with an asterisk (*)
are significantly different from random (binomial probability test, P < 0.05).
(Fig. 3). This contrast holds when comparisons
are made to ERG data taken from light-adapted
green turtles as well (Granda and O'Shea, 1972).
Mrosovsky (1972), using broad-band blue and
red light, also observed discrepancies between
behavioral responses of green turtles and the
action spectra provided by Granda and O'Shea.
Although evidence suggests that green turtle
hatchlings use some assessment of spectral qual-
ity in sea finding, behavioral experiments that
more closely match ERG methods are needed
to quantify this pattern.
A phototactic response positively biasing
short-wavelength light also has been found
among a variety of anurans (Hailman and Jae-
ger, 1974). Those workers have rejected the
contention of Muntz (1962) that the response
is an escape mechanism directly resulting in wa-
ter finding but instead propose an "open sky"
attraction hypothesis. In the case of hatchling
sea turtles, phototaxis with a spectral bias for
short-wavelength light, either reflected from a
blue ocean or scattered from a blue sky, may
result in seaward orientation under a majority
of conditions.
Although no action spectrum exists that would
allow a physiological-behavioral comparison to
be made for the loggerhead, the manner in
This content downloaded from 164.51.49.160 on Thu, 29 May 2014 13:28:38 PM
All use subject to JSTOR Terms and Conditions
WITHERINGTON AND BJORNDAL-SEA TURTLE PHOTOTAXIS 1067
-- BEHAVIORAL RESPONSE
----- ERG RESPONSE 60 MICROVOLTS
---- ERG RESPONSE 20 MICROVOLTS
1.0-
w
C') z 0
w
-j
0.0
300 400 500 600 700
WAVELENGTH (NM)
Fig. 5. Behavioral and action responses to spectral light in hatchling green turtles. Relative response for
the behavioral curve is the proportion of green turtle hatchlings choosing a light source varying among the
wavelengths specified at log intensity 3.5 (1.44 x 1017 photons s-' m-'), over a source of constant color and
intensity. Action response curves are adapted from Granda and O'Shea (1972) and represent log sensitivity
among wavelengths, as measured by electroretinography (ERG). Relative response for the ERG curves is
intensity sufficient to evoke 60 and 20 microvolt criterion responses. The high and low criterion response
curves represent high and low light intensity levels. Absolute intensity for the ERG light source was reported
to be 2.7 x 101~ photons s- m-2 at 580 nm.
which loggerhead hatchlings behave toward
spectral light indicates that they too may use
spectral cues in sea finding. Whereas logger-
head hatchlings orient positively to near-ultra-
violet, violet, green, and red light (Fig. 2), they
avoid yellow and yellow-orange light (Fig. 4a-
b). The aversion that loggerhead hatchlings
show toward yellow light, or xanthophobic re-
sponse, has also been observed in loggerhead
hatchlings orienting on a natural beach when
presented 590 nm monochromatic yellow light
from a low pressure sodium vapor (LPS) light
source (Witherington and Bjorndal, 1991). Re-
gardless of distance, no loggerhead hatchling
was attracted to the LPS light source. In one of
the first papers describing sea finding in sea
turtles, Hooker (1911) reported that logger-
head hatchlings on a beach during the day re-
sponded negatively to panes of glass transmit-
ting primarily orange-red light. Hooker could
not be certain, however, that this response in-
dicated a reaction to color. Responses identical
to those Hooker observed are predicted from
hatchlings exhibiting simple phototaxis. The
responses observed from loggerhead hatchlings
in the laboratory (present study) and on the
beach (Witherington and Bjorndal, 1991), how-
ever, strongly suggest a response to color. If the
behavior we observed was simply a response to
light of a specified brightness, the same response
would have been expected to wavelengths and
intensities other than the two intensities of yel-
low wavelengths (Fig. 4a-b).
A negative response to long-wavelength light
(576 and 605 nm) during escape behavior has
been reported for the leopard frog (Rana pi-
piens) (Fite et al., 1978). Because the normal
escape response of this frog is toward water,
this mechanism may share functional similari-
ties with that of the loggerhead. One function
This content downloaded from 164.51.49.160 on Thu, 29 May 2014 13:28:38 PM
All use subject to JSTOR Terms and Conditions
1068 COPEIA, 1991, NO. 4
of xanthophobia in loggerhead hatchlings may
be to reduce the attraction of light sources with
a substantial participation of long-wavelength
light. Such an adaptation could be advanta-
geous if long-wavelength light sources with the
potential to disrupt the sea-finding ability of a
directional or phototropotaxis mechanism were
common in nature. Rising and setting celestial
bodies appear as predominantly long-wave-
length sources because of the short-wavelength
scattering effect of the atmosphere and have
the potential to disrupt hatchling sea finding.
Some controversy exists regarding whether the
rising sun affects sea finding in sea turtles.
Whereas Ehrenfeld and Carr (1967) and van
Rhijn (1979) report that green turtles and
hawksbill turtles (Eretmochelys imbricata) are af-
fected insignificantly by the sun on the horizon,
Mrosovsky (1970) and Mrosovsky and Kingsmill
(1985) report that loggerhead, green, and
hawksbill turtles are significantly affected. The
loggerhead hatchlings in Mrosovsky's study ori-
enting at sunrise or sunset without the contrast
of a dune horizon still moved in the general
ocean direction. It is remarkable that the sun,
as an intense opposing light source, affected ori-
entation in these experiments as little as it did.
The positive response of loggerhead hatch-
lings to 700 nm light at high intensity shows
that a comprehensive bias against long-wave-
length light does not exist. In addition to being
on the periphery of color sensitivity in logger-
heads, 700 nm light may also fall outside their
ability for color discrimination. The variation
in response of loggerhead hatchlings to 540,
560, and 580 nm light at varying intensity in-
dicates that the xanthophobic response is not
independent of light intensity.
Models of mechanisms by which sea turtles
achieve a seaward orientation commonly em-
ploy the term brightness to denote the cue that
guides hatchlings to the ocean (Verheijen and
Wildschut, 1973; van Rhijn, 1979; Mrosovsky
and Kingsmill, 1985). Unfortunately, bright-
ness in this usage is not a currently measurable
value. Brightness from the perspective of the
sea turtle hatchling must certainly incorporate
intensity in proportion to a species-specific ac-
tion spectrum. Could perceived brightness,
however, be influenced by other biased re-
sponses to color? Brightest-direction models
must incorporate a definition for brightness that
considers such complexities if those models are
to explain the orientation behavior we observed
in loggerhead and green turtle hatchlings.
ACKNOWLEDGMENTS
We thank Florida Power and Light Company,
the National Fish and Wildlife Foundation, and
the United States Fish and Wildlife Service for
funding this study. We are especially grateful
to E. Possardt and R. Wilcox for their support.
J. Provancha and R. Wheeler of Bionetics Com-
pany provided the spectroradiometer in addi-
tion to technical counsel. L. Ehrhart graciously
allowed us access to his study area, and T. Hold-
erfield and W. Hancock aided in the field work.
A. Bolten provided invaluable advice through-
out the study, J. Brockmann and N. Mrosovsky
reviewed earlier drafts of the manuscript, and
J. Miranda aided in the construction of the ap-
paratus.
LITERATURE CITED
CARR, A., AND L. OGREN. 1960. The ecology and
migrations of sea turtles. IV. The green turtle in
the Caribbean sea. Amer. Mus. Nat. Hist. Bull. 121:
1-48.
DANIEL, R. S., AND K. U. SMITH. 1947. The sea-
approach behavior of the neonate loggerhead turtle
(Caretta caretta). J. Comp. Physiol. Psychol. 40:413-
420.
EHRENFELD, D. W. 1968. The role of vision in the
sea-finding orientation of the green turtle (Chelonia
mydas). II. Orientation mechanism and range of
spectral sensitivity. Anim. Behav. 16:281-287.
•-,
AND A. CARR. 1967. The role of vision in
the sea-finding orientation of the green turtle (Che-
lonia mydas). Ibid. 15:25-36.
FITE, K. V., J. SOUKUP, AND R. G. CAREY. 1978.
Wavelength discrimination in the leopard frog: a
reexamination. Brain Behav. Evol. 15:405-414.
GRANDA, A. M., AND P. J. O'SHEA. 1972. Spectral
sensitivity of the green turtle (Chelonia mydas mydas)
determined by electrical responses to heterochro-
matic light. Ibid. 5:143-154.
,, AND C. A. DVORAK. 1977. Vision in turtles,
p. 451-495. In: Handbook of sensory physiology.
Vol. VII/5, the visual system in vertebrates. F.
Crescitelli (ed.). Springer-Verlag, Berlin, Germany.
HAILMAN,J. P., AND R. G. JAEGER. 1974. Phototactic
responses to spectrally dominant stimuli and use of
colour vision by adult anuran amphibians: a com-
parative survey. Anim. Behav. 22:757-795.
HOOKER, D. 1911. Certain reactions to color in the
young loggerhead turtle. Carnegie Inst. Washing-
ton Publ. 132:71-76.
LIMPus, C. J. 1971. Sea turtle ocean finding behav-
iour. Search 2:385-387.
LOEW, E. R., ANDJ. N. LYTHGOE. 1985. The ecology
of colour vision. Endeavour 14:170-174.
MROSOVSKY, N. 1968. Nocturnal emergence of
This content downloaded from 164.51.49.160 on Thu, 29 May 2014 13:28:38 PM
All use subject to JSTOR Terms and Conditions
WITHERINGTON AND BJORNDAL-SEA TURTLE PHOTOTAXIS 1069
hatchling sea turtles: control by thermal inhibition
of activity. Nature 220:1338-1339.
-. 1970. The influence of the sun's position and
elevated cues on the orientation of hatchling sea
turtles. Anim. 'Behav. 18:648-651.
. 1972. The water-finding ability of sea turtles.
Brain Behav. Evol. 5:202-225.
-, AND A. CARR. 1967. Preference for light of
short wavelengths in hatchling green sea turtles,
Chelonia mydas, tested on their natural nesting
beaches. Behaviour 28:217-231.
~, and S. J. Shettleworth. 1968. Wavelength
preferences and brightness cues in the water find-
ing behaviour of sea turtles. Ibid. 32:211-257.
- , AND S. F. KINGSMILL. 1985. How turtles find
the sea. Z. Tierpsychol. 67:237-256.
MUNTZ, W. R. A. 1962. Effectiveness of different
colors of light in releasing positive phototactic be-
havior of frogs, and a possible function of the ret-
inal projection to the diencephalon. J. Neurophysi-
ol. 25:712-720.
PARKER, G. H. 1922. The crawling of young logger-
head turtle toward the sea. J. Exp. Zool. 6:323-
331.
VAN RHIJN, F. A. 1979. Optic orientation in hatch-
lings of the sea turtle, Chelonia mydas. I. Brightness:
not the only optic cue in sea-finding orientation.
Mar. Behav. Physiol. 6:105-121.
~I, AND J. C. VAN GORKOM. 1983. Optic orien-
tation in hatchlings of the sea turtle, Chelonia mydas.
III. Sea-finding behaviour: the role of photic and
visual orientation in animals walking on the spot
under laboratory conditions. Ibid. 9:211-228.
VERHEIJEN, F. J., AND J. T. WILDSCHUT. 1973. The
photic orientation of sea turtles during water find-
ing behaviour. Neth. J. Sea Res. 7:53-67.
WITHERINGTON, B. E., AND K. A. BJORNDAL. 1991.
Influences of artificial lighting on the seaward ori-
entation of hatchling loggerhead turtles (Caretta ca-
retta). Biol. Conserv. 55:139-149.
, -, AND C. M. MCCABE. 1990. Temporal
pattern of nocturnal emergence of loggerhead turtle
hatchlings from natural nests. Copeia 1990:1165-
1168.
ARCHIE CARR CENTER FOR SEA TURTLE RE-
SEARCH AND DEPARTMENT OF ZOOLOGY,
UNIVERSITY OF FLORIDA, GAINESVILLE,
FLORIDA 32611. Accepted 19 Nov. 1990.
This content downloaded from 164.51.49.160 on Thu, 29 May 2014 13:28:38 PM
All use subject to JSTOR Terms and Conditions