C
THE PATTERN CHANGE OF ACCLIMATION TO NOVEL GRAINED SUBSTRATES
OF THE SPECKLED SANDDAB
ABSTRACT
studied the pattern changing abilities of the Speckled Sanddab by
acclimating sandy patterned fish to a rocky substrate and rocky
patterned fish to a sandy substrate. Both rocky and sandy fish
demonstrated the ability to alter their pattern to match a nove
substrate. After 8 days of acclimation, rocky fish matched sand bettet
than sandy fish matched rock. Whether this was the result of all fish
becoming sandier af ter removal from the field is not certain. There may
be a cost to maintaining a rocky morphology. If so, it would be more
difficult for a sandy fish to gain a rocky morphology than it would for a
rocky fish to gain a sandy morphology. Sandy fish showed a greater
fidelity to their original substrate than rocky fish, indicating that a
lower ability to effectively match another substrate limits the freedom
afish displays in choosing a new substrate.
INTRODUCTION
Cryptic coloration is an important method of defense for many
animals. The ability to match a background effectively in the presence
of visual predators can be cructal for survival. Flatfish are among the
best camouf laged of all fish, with their ability to lie flat on the ocean
floor buried in the sand or mud. Ideally, a species using cryptic
coloration as a survival tool should be able to match any substrate on
which it is placed, yet most creatures naturally selected to match a
certain environment are limited in their abilities to adapt to novel
environments
while SCUBA diving in the Hopkins Harine Life Refuge, 1 observed that
the sanddabs, Citharichthys stigmaeus, on a uniform sand bottom have
extremely different pigment patterns than the sanddabs on the adjacent
substrate of gravel, broken shells, and algal fragments closer to shore
This led me to ask two questions: How well and how quickly can
sanddabs match new backgrounds? Do these fish remain on the substrate
to which they are best adapted, or do they move about freely between
substrates?
Research on flatfish coloration shows that most species of flatfish
have the ability to change or adjust their color patterns using two
methods. There is an instantaneous color change, which is an immediate
and temporary lightening or darkening of the fish, and there is a
morphological change, which takes place over a longer period of time
resulting in a more permanent change in color or pattern. Conditions that
cause instantaneous pattern changes eventually lead to morphological
pattern changes (Brown, 1957.) Studies on plaice show that the young
fish, which are found nearshore, have semi-permanent white spots that
are not seen on adults, which live in the muddy bottom. However, when
the adult fish are placed on white pebbbles, these spots reappear
(Cott,1940.) Work with Paralichthys albiguttus demonstrates that this
fish can appear finely mottled when placed on a checkerboard of small
black and white squares and coarsely mottled on a checkerboard of large
squares (Parker, 1971.)
designed a set of experiments to answer my questions about the
background matching abilities and substrate fidelity of the Speckled
Sanddab. By placing different patterned fish on novel substrates,
observed that both rocky and sandy sanddabs possessed the ability to
alter their patterns to match new backgrounds, and that sandy fish
showed a grester fidelity to their substrate than rocky fish.
HATERIALS AND FETHODS
Using SCUBA gear and a small handnet, I collected 28 Speckled
Sanddabs, Citharichthys stigmaeus, of sizes 6.5 to 11 cm. Fourteen of
these fish had sandy patterns and were collected from the sandy
substrate in the channel between Hopkins Harine Station and the
Monterey Bay Aquarium at a depth of 20 feet. The other 14 fish were
collected at a depth of 10 feet from the area between Seal Rock and the
railroad tracks off Agassiz Beach, where the substrate consisted of
gravel, broken shells, sandy patches, and algae fragments; these fish had
speckled morphologies (for map of collection sites, see Figure 1.) Since
there was within site vartability of substrate type, especially within my
designated type "rocky", I made 10 measurements with a 25 cm x 25 cm
transect at each of the two collection sites to quantify the makeup of
the substrates. The transect was haphazardly dropped and substrate
descriptions were recorded from the four corners and the center.
Fish were anesthetized with quinaldine and tagged with a color code
of embroidery beads sewed through the posterior dorsal muscle mass
with nylon thread.
Fish were maintained in two outdoor tanks with equal exposure to
sünlight and fresh running seawater, each containing a substrate
collected from the two capture sites. Each 1.5 square meter tank
contained 7 experimental and 7 control fish. The fish were fed mussels
and squid every 4 days.
photographed each fish 8 hours after collection to record initial
pattern morphology. The fish were then placed in the tanks and
photographed 24 hours after initial acclimation and every 48 hours
thereafter to document the pattern change as adaptation to a different
substrate occurred. For photographing, fish were placed in a dish of
seawater with a gray scale and ruler as a background. They were
Iluminated with four 300 watt reflector flood lamps and photographed
using Agfa Pan 100 speed film in a 35 FM Nikon F2 camera mounted 22
cin above the dish. Photographs were also taken of the two substrates,
Fish were tested for substrate preference in an aquarium in which
the bottom was half sand and half rock. The substrates in the tank were
collected from the fish capture sites. Fish were quided onto the dividing
line between rock and sand through a clear cylindrical chute. Each
control fish was sent down the chute 3 times and substrate choices were
scored. This was repeated for each control fish every day for 17 days.
Two methods of analyzing pattern changes were employed: human
ranking and digital image processing. A pattern scale of 1 (most rocky)
to 5 (most sandy) was established based on five representative
photographs of different fish (see Figure 2). 25 photographs of fish at
day O and 25 photographs of the same fish acclimated for 8 days were
shuffled together and scored on this scale of 1 to 5 by 15 unbjased
judges. Three sandy patterned fish acclimating to rock died during the 8
day period and could not be scored. By subtracting the pre-acclimation
score from the post-acclimation score for each fish, l obtained a value
for the magnitude and direction of pattern change. A positive change
indicates the fish became more sandy; a negative change indicates the
fish became more rocky. Hean pre- and post-acclimation values, as well
as net change values, were obtained for each group of fish (rocky and
sandy, controls and experimentals.)
he rate of acclimation was determined from 6 photographs taken
during the acclimation of 4 fish. The middle 4 photographs in each series
were removed and shuffled, leaving the first and last photographs as
endpoints. 5 unbiased judges were instructed to place the 4
intermediate photographs from each series in order from initial pattern
morphology to final. Each misplaced photograph was assigned a value of
to 3, depending on how far from its correct position the photograph
was placed (for example; if the series' original order was "ABCD" and the
photographs were arranged "BDCA", the A position would be assigned 1.
the B position 2, the C position O, and the D position 3.) By observing the
time after which visually obvious changes in pattern morphology could no
longer be distinguished, I was able to estimate sanddab acclimation rate
to a new substrate
Digital image processing was used to analyze a few representative
photographs as a demonstration that this technology may be applied as a
valuable complement to other forms of photographic analysis. The image
processing was conducted on the MegaVision 1024-XM computer with
input from a video camera. Photographic images were stored in 1024 X
1024 arrays of pixels, with each pixel assigned a gray value from 0
(black) to 255 (white).I selected pre- and post-acclimation photos for
one rocky fish placed on sand and for one sandy fish placed on rock (see
Figures 4 & 5, respectively), Histograms of the gray scale values for a
small area around the same eyespot on each fish were obtained to
compare how this specific area changed af ter 8 days of acclimation to a
different substrate. Histograms of gray scale values for equal sized
areas of sand and rock were also obtained for comparison of the two
substrates. The command "scanner" was used to input the video image of
the black and white print into one of the temporary memories. "Polygon
was used to create a template so that several photographs of the same
fish could be aligned. This command was then used to create a sample
area for the analysis of gray scale values from an image. "Hpolygon
created a histogram of gray scale values from the image within the
polygon. "Hlist" was used to obtain the raw data from the histogram,
which was stored onto a floppy disc and transferred to Lotus for
graphing. To compensate for differences in exposure or development of
the photographs, "sample" was used to obtain point readings of gray scale
values from the image, and "mapper" was used to equalize existing
differences.
RESULTS
Pre-acclimation rocky fish had an average pattern score of 1.75,
whereas pre-acclimation sandy fish had an average pattern score of 4.48.
After the 8 day acclimation, rocky fish on rock became sandter by a
degree of 1.04; rocky fish on sand became sandier by 2.41 degrees. The
sandy fish on sand showed an increase in sandiness of 46 degrees,
whereas the sandy fish on rock became rockier by 53 degrees (See Table
and Figures 4-8). One-tailed, rank sum two sample tests show that
there is a significant difference between the pattern change of
experimental and control fish (p-.006 for rocky fish, p-03 for sandy
fish, see Table 1.) After 8 days of acclimation, rocky fish matched a
sandy substrate better than sandy fish matched a rocky substrate (See
Figures 9 & 10.)
When given the choice between rocky and sandy substrates, rocky
control fish chose the sand 56% of the trials. Sandy control fish
demonstrated a greater preference for sand by choosing the sandy
substrate 76% of the trials (See Table 2 and Figure 11) A Mann-whitney
statistical analysis was performed between the means of percent
preference for sandy substrate by rocky and sandy controls to
demonstrate that this difference is significant (p-.0025, see Table 2.)
Most of the visually obvious pattern changes of sanddabs acclimating
to a new substrate took place within the first 3 to 5 days. It was this
time range in which it became difficult to distinguish the differences
between photographs of the same fish (See Table 3.)
At the collection site for the sandy fish, the substrate consisted of
100% sand. However, at the collection site for the rocky fish, the
substrate was composed of 488 rocky gravel and broken shells coyered
with algae; 428 bare, finer gravel, and 102 sand, which was dispersed in
small patches ranging from 5 to 10 inches in diameter.
The histogram of rocky substrate shows a much greater spread of
gray values than the histogram of sandy substrate. The histograms
documenting the change of gray values of the rocky fish before and after
acclimation to sand show that the spread of gray values is greatly
reduced after 8 days. The histograms of the sandy fish acclimating to
rock show little change except a general darkening (see Figures 12-15.)
DISCUSSION
Flatfish possess a large ability to change their patterns to adapt to
new substrates. The pattern is largely determined by the percent cover
within independent patch distributions of two different pigment
structures found in the speckled sanddab, melanophores and iridophores.
The melanophore is a reticulate cell that can disperse or concentrate the
dark pigment it contains, quickly varying its contribution to the
background shade. The iridophore contains intracellular layers of white
quanine crystals which are fixed in position. Iridophores can not quickly
vary their contribution to background shade like melanophores, (Hoar,
1969.) In an instantaneous change, the pattern and degree to which
existing melanophores are open is varied against the fixed background
pattern of tridophores. A morphological pattern change results in a
change of the total amount of pigment or net number of pigment cells
present. Morphological changes occur within melanophore patterns, but
may also occur more slowly with iridophore patterns. Acclimations
longer than 8 days would need to be performed to observe changes in the
patterns of tridophores.
The patterns of rocky and sandy fish are extremely different, yet
both types express some aspect of a basic pattern of black eyespots and
white dorsal, ventral, and lateral line spots. Observations through a
dissecting scope show that the black eyespots are composed mainly of
melanophores and the white spots are composed mainly of iridophores.
On a rocky fish, tridophores are densely packed to form the
characteristic white spots and are sparsely distributed over the rest of
the fish. On a sandy fish, these iridophores are much more evenly
distributed, with only loose clusters of iridophores at the characteristic
spot sites. On the rocky fish, the melanophores highlight the white
irridophore spots by being densely packed in the areas between the spots
and by being sparsely distributed over the spots. On the sandy fish, the
melanophores are evenly distributed among the iridophores. It is not
exactly clear whether these areas of densely packed pigment structures
contain a high density of cells, a high concentration of pigments, or both.
It is also not clear whether areas of sparse distibution of pigment
structures contain a low density of cells, a low concentration of
pigments, or a combination of the two.
When a rocky fish acclimates to sand, it lightens the dark areas of
its pattern and darkens the white spots to achieve an even tone. This can
be seen in the histograms of the rocky fish before and after acclimation
to sand. The histogram before the acclimation shows a wide spread of
gray values, whereas the histogram af ter the acclimation shows a much
narrower range of values, much like the histogram of the sandy substrate
(see Figure 14) The fish, however, does not lose its blotchy pattern,
only its contrast. This suggests that the pattern change of the 8 day
acclimation was a change in the melanophores, and not a rearrangement
of the iridophores. Whether the fish is actually decreasing its total
number of melanophores in the dark areas and increasing the number of
melanophores over the iridophore spots, or is merely adjusting existing
melanophores is not certain. It is also not clear whether the irridophore
pattern is completely permanent or if it can show plasticity after
acclimation longer than 8 days.
Asandy fish acclimating to rock shows the reverse type of response
of a rocky fish acclimating to sand. The iridophore spots become lighter
and the area between the spots becomes darker. Although contrast was
changed, the sandy pattern was never lost, again suggesting a change in
the melanophores rather than in the iridophores. However, since
fridophores are less densely packed in the pattern spots, the lightened
areas are not as large or white as those of a rocky fish. The histograms
of the sandy fish acclimating to rock Illustrate this point by showing
little change in the spread of gray values af ter the acclimation and show
only a general darkening (see Figure 15.)
It is interesting that there is a basic pattern expressed by all
sanddabs. This pattern, which can be used so effectively to adapt to new
substrates by increasing and decreasing contrast, gives any sanddab that
lands on a coarse graîned substrate a chance to match adequate ly until
the final pattern can be perfected, if the fish remains on that substrate.
Also, this pattern does not interfere with matching to a fine grained
substrate. The pattern is advantageous for a rocky fish because rocky
substrate is not 100% rock; it is composed of rock, gravel, broken
shells, and sand patches. Fish that possess this pattern can adapt to all
of these quickly as they move about the rocky substrate and are likely to
be selected for. Sandy patterned fish possess aspects of this easily
adaptible pattern as well, showing that perhaps these fish are migratory
and move over a variety of substrates, or that their life cycle may
expose them to different grains of substrates at different stages of
development. The grain of a sandy substrate to a tiny, newly-settled
fish may be relatively large, so the adaptable pattern set on a small
scale serves the tiny fish well.
Since rocky fish acclimating to sand matched the sand better than
sandy fish acclimating to rock matched the rock, it would seem that
rocky fish show a much greater ability to change than sandy fish (See
Figures 9 ≈ 10) However, there is a general trend for the fish in my
experiment to become more sandy. Both control groups displayed an
increase in sandiness af ter 8 days. This could be the result of one or a
C
combination of several factors. Perhaps the substrates in both tanks
were actually "sandter" than the substrates in the field. Overcrowding or
the dict of oquid and musoclo may have pomchow decreased rocky pattern
stability. Since sandy fish were collected 10 feet deeper than the rocky
fish, and light decreases with ocean depth, perhaps a decrease in light
causes an increase in sandiness. If this were true, and the fish in the
tanks recieved less light than fish at 20 feet of depth, sandiness would
be increased.
Another explanation is that there is some cost involved with
possessing a rocky morphology, and that removing these fish from the
field somehow reduced the pressure to maintain this morphology, causing
the fish to become more sandy. If this is the case, then this cost is most
likely the energy spent on maintaining the expression of the highly
plastic melanophores in dark bands and patches.
Sandy fish, when given the choice between rock and sand, seemed to
display a preference for sand by choosing sand 768 of the trials. Rocky
fish, when presented with the same choice, displayed little preference
for either rock or sand. This correlates well with what has been
presented. Sandy fish do not seem able to immediately match a rocky
substrate. This could be the result of not having large irridophore spots
to higfilighit with melanophore changes and create a high contrast
pattern, or it could be that obtaining and maintaining a rocky morphology
is too energetically costly. Since sandy fish are not able to bury in the
rocky substrate for camof lage, they remain in the sand where they match
best. Rocky fish, however, seem to be able to adapt quickly to a sandy
substrate. This could be the result of possessing an irridophore pattern
with easily reduceable contrast, or it could be energetically favorable to
discontinue maintenance of melanophore bands and patches. Fish
acclimating to a sandy substrate also have the ability to bury in the sand
for added camouf lage during the early stages of adaptation, since it
takes 3 to S days to complete the matching. Rocky fish do not seem to be
limited by an insbility to background match when choosing between rock
and sand.
If à cost to maintaining a rocky morphology does exist, then what
sort of benefits does a rocky fish recieve to counteract that cost? It
allows the fish to to exploit a habitat unavailable to their most
efffective competitors, members of the same species adapted to sandy
substrate. Perhaps Iiving on a rocky substrate provides access to better
food. More or tastier crustaceans may live within the gravel and broken
shells than in the sandy bottom. The sanddab may eat the algae that
settles or grows only on the rocky substrate. Perhaps the advantage is
less exposure to larger open water predators that may not come into the
shallow waters nearshore.
CONCLUSION
Rocky and sandy sanddabs both demonstrated an ability to change
their patterns to adapt to novel substrates. Rocky fish matched sand
better after 8 days of acclimation than sandy fish matched rock, which
was probably the result of the general trend for the fish in the
experiment to become more sandy. All fish displayed aspects of the
same basic pattern of iridophore spots, although the pattern was more
pronounced in the rocky fish. Since it is eastly highlighted or
de-accentuated, probably by the action of melanophores, this basic
pattern seems to offer a great deal of flexibility to the sanddab in
matching many grains of substrates.
C
Literature Cited
Brown, Hargaret E. 1957. The Physiology of Eishes, Volume 111.
Academic Press Inc. pp 387-401.
Cott, Hugh B. 1940. Adaptive Coloration in Animals. Oxford University
Press. pp 27-30.
Hoar, W. S. and D. J. Randall. 1969. Fish Physiology, Volume 111.
Academic Press Inc. pp 307-313.
Parker, George Howard, 1971. Animal Color Changes and Their
Neurobumours. Hafner Publishing Co. pp 169-175.0
Pasternak, Dahna. 1987. "Quantitative Analysis of Background Matching
of Sepia officinalis" Stanford University Hopkins Marine Station Final
Papers Biology 175H. (unpublished.)
Zar, Jerrold H. 1974. Biostatistical Analysis. Prentice-Hall Inc. pp
09-12
ACKNOWLEDGEMENTS: Hopkins Marine Station in Pacif ic Grove California
is à poem, a stink, à grating noise, a quality of light, a tone, a habit, a
nostalgia, a dream. Hopkins Marine Station is the gathered and scattered,
the pyrex dish, the plastic bucket, the broken gate and weedy lots and
hidden keys, sanddab fishtanks of corrugated iron, honky tonks(Agassiz),
restaurants(JC canyon), and whore houses(Hawthorne), and little crowded
groceries(Nob Hill), and laboratories(Lily), and flophouses(Fisher). Its
inhabitants are, as Chuck Baxter once said, "whores, pimps, gamblers,
and sons of bitches, by which he meant the Spring Class. Had the man
looked through another peephole he might have said, "Saints and angels
and martyrs and holy men of science," and he would have meant the same
thing.
would like to thank Chuck Baxter, my advisor, for the exchange of ideas,
be they in the form of taxonomic loquacities, insightful paper revisions,
or hasty pool shooting blather.
would like to thank Mark Denny, my secondary advisor, for all of his
help, advice, assistance, chicken curry, and fourrier transform
explanations.
would like to thank Ladd Johnson for taking some large chunks of time
to really sit (split infinitive) down and look at the data. l'Il name my
first ANOVA after you.
Iwould like to thank the Squid Festival for forcing Emily Carrington to
seek out strange new buddy forms for some amazing diversions. Here we
are on ze Calypso...
would like to thank Chris Patton for film, for opening up the door, and
for film.
And finally, would like to thank the words "rocky", "sandy", "acclimate"
and "substrate", without whose existence this paper would not have been
possible.
O
Table Legends
Table 1. Mean pattern morphology values for each fish day O
day 8, and
the net change. Below: one tailed, rank sum two sample statistical
sts.
Table 2. Percent selection of sandy substrate by rocky and sandy
patterned fish. Below: Mann-Whitney statistical test.
Table 3. Rate of acclimation determination: values assinged to
misplaced photographs.
O
ROt
CONTROL
(on reck)
ROCKY FISH
XFERIMI
ENTAL
(on sand)
SANDY FISI
CUNTROL
(on sand)
SANDY FI
RIMENTAL
(an reck)
(ONE TAILE
FIEH
ROCI
SANDY FISH
RAHE
TAELE 1.
INDIVTDUAL
MEAN
D.DEV.
SAMFLE
ERRUR
STD.DEV.
SAMFLE
D.ERROE
FEA
TD.DE
SAMFLE S
STD.E
FiEAN
TD. DEV.
SAMFLE
SUM TUO
CHOMGE
CHANG
CHANC
CHANGE OF
SAMFLET
CONTRCI
EXFERI
ENTAL
RO
XFERIHENTAL
FAT
Before
1.1
1.6
3
1
4
1.76
O.33
6.27
.1
1.
1.53
1.C
1.
1.
O..
O.18
4.

367
4.8
41
4.8
4.41
O.C
0.24
2.6
473
4.73
4.7
0.74
O
MORFHOLOE
After
93
2.1
87
1.8
3.8

0.77
O.E
3.7.
433
3.6
3.7
3. 74
O.
0.08
4.33
47
4.
4.93
4.87
0.
0,08
3.6
48
4.67
4.O
0.38
N=7
=7
=7
=
VALUE
Change
6.73
1.
1.6
1 „07
0.54
O6
133
1.04
0.41
O.13
1.8



41
0.6

0.
1.
0.13
O.13
0.2
0.46
O.6
0.2
-04
—1.3
-0.13
-O.26
-O.
O.2
Fz00
AELE.

Terce
Sandy Subs
Racky
Centrals
7.144
66.67
1.7
7.6
a6

41.6
1.6
3.3
5.41
ear
19
Standard Deviation
Sample Size
andard

MH-WHITNEY STATISTICAL
EY COMTROL
7 SAMDY SUESTRATE
CHOSEN
=18
FOO
ROCKY COLOR
M=16
SANDV
COLORED
Sand
ant
.
0
76.19
76.1.

S0
76.17
C71
70.48
71.4
76.19
7.14
17.6
78
77
61.11
18
266
ted
O
ISFLACEI
50
TRIAL.
1

VALUE
ASSICN

HOTOS
O MISFLALED


.
1
1
1
1


1
Figure legends
Figure 1. Map of Speckled Sanddab collection sites in Hopkins Marine
Life Refuge.
Figure 2. Scale of pattern morphology. 1-most rocky, 5-most sandy.
Figure 3. Speckled Sanddab pattern of irridophore spots. Circle
indicates area sampled by eyespot histograms of gray values.
Figure 4. Change in pattern of rocky fish acclimating to sand for 8 days.
Figure 5. Change in pattern of sandy fish acclimating to rock for 8 days.
Figure 6. Change in pattern of rocky fish acclimating to rock for 8 days.
Figure 7. Change in pattern of sandy fish acclimating to sand for 8
days.
Figure 8. Quantification of pattern changes of rocky and sandy fish,
based on scale in Figure 2. (Error bars indicate +/- one standard
error.)
Figure 9. Substrate matching of rocky fish acclimated to sand for 12
days compared to sandy fish.
Figure 10. Substrate matching of sandy fish acclimated to rock for 12
days compared to rocky fish.
Figure 11. Substrate preference of rocky and sandy fish (4/- one
standard error.)
Figure 12. Histogram of gray scale values of rocky substrate.
Figure 13. Histogram of gray scale values of sandy substrate.
Figure 14. Histogram of gray scale values of rocky fish eyespot (see
figure 3), O days and 8 days acclimation to sand.
Figure 15. Histogram of gray scale values of sandy fish eyespot, O days
and 8 days acclimation to rock.
DD
8

.





V


Sa
6

5

—






.

:
od
O
Fiquse 3.

2
. .
Figure 4
Bockeg Fieh on Hand br 8 Dags

5112
O CUV2

Befoce
FURNITURE
« CHEMICALS

uilli
Afrec
1ina

ili
ig
V
Sci
10
iquce 5.
Saalh Fish da Rock 6. 8 dags
Oc
oc
1
3/2
AM
B212
tttitttittst tit t t rtotrtrt rt trtrtrt tittf

Before
u

Afre¬
Fiquee b.
Boclg Fish on Rock fe 8 Dags

.
IVOW
H
JOIIN
RAs
t tit
Firtrtrtrtrtrt t  .
Befoce
f

00

M
Af+ec
Hguce 7.
Sandg Fish on Saad for 8 Dags
c
5/12
HMA
661
Be foce
1
20
5
MAT
Afre¬
SO
80II
f
SIVOIW3HD
JaOLINan
o
E::
...*. .
. . . .
—

O


—
J
o
8
se
2700
2
).
5
P
5h5
te
C



1
1
1
1L1
+

II1


UEU EE Kpurg juemag
1
I

O
0
8
1

O
A

1
(spupsnou)
sexia jo aqun

o
- +
0 -
1

O
O

1



9
(spupsnoul)
sexia jo equnN
9
D,
O
1
1
I
O
1


V





—





—
—

oooo
(spupsnou)
sexia jo qun
+

0
O
1

I

1
O
2


O

O




ao

- —

ooo
(spupsnoul)
sexi jo qun
3