Ontogenetic change in novel functions: waterfall climbing

UNCORRECTED PROOF

Ontogenetic change in novel functions: waterfall climbingin adult Hawaiian gobiid fishes

R. W. Blob1, K. M. Wright1, M. Becker2, T. Maie1, T. J. Iverson2, M. L. Julius2 & H. L. Schoenfuss2

1 Department of Biological Sciences, Clemson University, Clemson, SC, USA

2 Department of Biological Sciences, St. Cloud State University, Math and Science Center, St Cloud, MN, USA

Keywords

locomotion; performance; kinematics;

ontogeny; climbing; biomechanics; fish;

muscle.

Correspondence

Richard W. Blob, Department of Biological

Sciences, Clemson University, 132 Long

Hall, Clemson, SC, 29634, USA.

Tel: + 1 864 656 3602;

Fax: + 1 864 656 0435

Email: [email protected]

Received 31 October 2006; accepted

19 January 2007

doi:10.1111/j.1469-7998.2007.00315.x

Abstract

Juveniles from three species of Hawaiian gobiid fish climb waterfalls as part of an

amphidromous life cycle, allowing them to re-penetrate adult upstream habitats

after being swept out to the ocean upon hatching. The importance of climbing for

juvenile stream gobies is well established, but adult fish in upstream island habitats

also face potential downstream displacement by periodic disturbances. Thus,

retention of climbing ability could be advantageous for adult stream gobies.

Climbing performance might be expected to decline among adults, however, due

to the tendency for mass-specific muscular power production to decrease with

body size, and a lack of positively allometric growth among structures like the

pelvic sucker that support body weight against gravity. To evaluate changes in

waterfall-climbing ability with body size in Hawaiian stream gobies, we compared

the climbing performance and kinematics between adults and juveniles from three

species Awaous guamensis, Sicyopterus stimpsoni and Lentipes concolor. For

species in which juveniles climbed using ‘powerbursts’ of axial undulation, adult

performance and kinematics showed marked changes: adult A. guamensis failed to

climb, and adult L. concolor used multiple pectoral fin adductions to crutch up

surfaces at slow speeds, rather than rapid powerbursts. Adult S. stimpsoni, like

juveniles, still used oral and pelvic suckers to ‘inch’ up surfaces and climbed

at speeds comparable to those of juveniles. However, unlike juveniles, adult

S. stimpsoni also add pectoral fin crutching to every climbing cycle. Thus, although

powerburst species appear to be particularly susceptible to size-related declines in

waterfall-climbing performance, the addition of compensatory mechanisms pre-

vents the loss of this novel function in some species.

Introduction

Physical forces of the external environment place many

functional demands on animals. As animals grow, these

forces may change as a function of an animal’s increase in

size. To accommodate growth-related changes in external

forces, animals often exhibit compensatory mechanisms

such as allometric growth of support or propulsive struc-

tures (Carrier, 1996; Toro et al., 2003), or changes in

behavior (Higham et al., 2005), that help maintain func-

tional performance at a comparable level throughout

growth. Without such mechanisms, the performance of

some functions could decrease as juveniles mature into

adults, even if those functions continued to be advantageous

among larger individuals (Maie, Schoenfuss & Blob, in

review).

The freshwater fauna of small volcanic islands encounter

particularly challenging environments. Streams on these

islands are commonly subject to flash floods, high-velocity

flows and waterfalls that obstruct access to upstream

reaches (Fitzsimons & Nishimoto, 1995; Fig. 1a). Although

the physical forces exerted by these hazards can be rigorous

for stream-dwelling animals, many species have evolved

novel functional capacities that help meet such demands.

For example, in the Hawaiian Islands, the five native species

of stream fishes (four gobies and one eleotrid) hatch in

freshwater but are swept by currents to the ocean, where

they share an oceanic larval phase for several months before

returning to freshwater as juveniles (Ego, 1956; Julius, Blob

& Schoenfuss, 2005). However, three of these species face

further challenges because their adult habitats are in up-

stream reaches that can be obstructed by waterfalls up to

thousands of body lengths tall (Fitzsimons & Nishimoto,

1995). Juveniles of these species use two different mechan-

isms to climb waterfalls (Nishimoto & Kuamo’o, 1992;

Schoenfuss & Blob, 2003). Juvenile Awaous guamensis (Fig.

1b) and Lentipes concolor (Fig. 1c) climb using rapid ‘power-

bursts’ of axial undulation initiated by pectoral fin adduc-

tion. These are followed by rest periods during which fish

adhere to the substrate using a ventral sucking disk formed

from fused pelvic fins. In contrast, the third species, Sicyop-

terus stimpsoni (Fig. 1d), slowly inches up waterfalls using

J Z O 3 1 5 B Dispatch: 5.3.07 Journal: JZO CE: Chandrika

Journal Name Manuscript No. Author Received: No. of pages: 10 TE: Premalatha Op: gsravi

Journal of Zoology (2007) c� 2007 The Authors. Journal compilation c� 2007 The Zoological Society of London 1

Journal of Zoology. Print ISSN 0952-8369

JZO 315(BW

UK

JZ

O 3

15.P

DF

05-M

ar-0

7 19

:57

9547

12 B

ytes

10

PAG

ES

n op

erat

or=

Rav

isha

nkar

)

UNCORRECTED PROOFthe pelvic disk and a second, oral disk with little or no fin

motion or axial undulation. The use of these mechanisms

allows repopulation of upstream habitats by oceanic larvae

after disturbances such as floods or hurricanes (Fitzsimons

& Nishimoto, 1995).

Although the importance of climbing to juvenile

A. guamensis, L. concolor and S. stimpsoni has been estab-

lished, adult fish in upstream Hawaiian habitats face many

similar risks, including displacement by catastrophic distur-

bances (Fitzsimons & Nishimoto, 1995). As a result, the

retention of climbing ability could be advantageous for

adult fish. However, as gobies increase in size, their ability

to climb waterfalls might decline for several reasons. First,

because mass-specific power production tends to decrease

with body size (Irschick et al., 2003), size-related limits to

power production might restrict the maximum body size at

which ‘powerbursting’ is a viable climbing mechanism (Blob

et al., 2006). This might limit climbing ability in adult

A. guamensis and L. concolor, or force these species to adopt

different climbing mechanics. Power demands might not

limit climbing by S. stimpsoni because of the slow cycle

frequencies of inching (Schoenfuss & Blob, 2003). However,

studies of ontogenetic scaling of body proportions in gobies

(including S. stimpsoni) have shown that the area of the

pelvic sucking disk grows isometrically with body length (i.e.

scales / BL2) and, thus, does not keep pace with increases

in the mass of the body (/ BL3) that it must keep attached

to the substrate (Maie et al., in review). As a result, even

inching climbers like S. stimpsoni might show reduced

climbing performance as they grow larger.

In this study, we test for size-related changes in climbing

performance and behavior by evaluating the climbing

of adult stream gobies (A. guamensis, L. concolor and

S. stimpsoni) and comparing their performance and kine-

matics with previous data (Schoenfuss & Blob, 2003; Blob

et al., 2006) from juveniles. Because of possible limits to

muscular power production and isometric scaling of the

adhesive sucker, we expect adult climbing performance to

decrease relative to that of conspecific juveniles. These

changes might be especially dramatic for powerburst species

because their high locomotor frequencies (Schoenfuss &

Blob, 2003) are disadvantageous for the high power produc-

tion required for climbing by larger fishes. Thus, species that

use powerburst climbing as juveniles might require major

changes in locomotor behavior in order to climb as adults.

Materials and methods

During three field seasons (2003–2005), adult climbing

gobies were collected with ‘opae nets’ (bowl-shaped mesh

nets with a narrow mouth) from habitats above waterfalls

in mid- (A. guamensis and S. stimpsoni) or upstream

(L. concolor) reaches of Hakalau Stream on the Island of

Hawai’i. Within 2 h of capture, animals were transferred in

stream water to a facility provided by the Hawai’i Depart-

ment of Land and Natural Resources, Division of Aquatic

Resources (DAR). Animals were separated by species into

groups of three to five individuals and acclimated together

for several hours at ambient stream temperature (19 1C).

Trials were conducted on individuals that overlapped in

body size (Table 1).

Our climbing arena (Fig. 2) was a chute that extended up

from a catch basin (Blob et al., 2006). The basin was 60 cm

long� 45 cm wide� 15 cm deep, half filled with Hakalau

stream water, and contained a rock to provide cover for the

fish. Chutes extended from the basin at 571 from horizontal,

and consisted of 1.5m long sections of a 4 in. diameter PVC

pipe (cut in half longitudinally) with fine-grained sand

Figure 1 (a) Falls of Nanue Stream on the Hamakua Coast of the island of Hawai’i, a waterfall climbed by juveniles of all three species of climbing

Hawaiian stream gobies: powerburst climbing Awaous guamensis and Lentipes concolor, and the inching climber Sicyopterus stimpsoni. (b)

Photo of adult A. guamensis. (c) Photo of adult male L. concolor. (d) Photo of adult male S. stimpsoni. Photos in parts b–d were provided courtesy

of Mike Yamamoto; all fish were free-swimming wild adults.

Journal of Zoology (2007) c� 2007 The Authors. Journal compilation c� 2007 The Zoological Society of London2

Ontogeny of waterfall climbing in gobiid fishes R. W. Blob et al.

JZO 315(BW

UK

JZ

O 3

15.P

DF

05-M

ar-0

7 19

:57

9547

12 B

ytes

10

PAG

ES

n op

erat

or=

Rav

isha

nkar

)

UNCORRECTED PROOF

Tab

le1

Mean

valu

es

and

sta

ndard

devia

tions

of

kin

em

atic

and

perf

orm

ance

variable

sfo

rju

venile

sand

adults

of

two

specie

sof

wate

rfall-

clim

bin

gH

aw

aiia

nstr

eam

gobie

s

Sic

yopte

rus

stim

psoni(S

s)

Lentipes

concolo

r(L

c)

Ss

adult

vers

us

juvenile

Lc

adult

vers

us

juvenile

Adult

Ss

vers

us

Lc

Adult

Juvenile

Adult

Juvenile

Fis

hle

ngth

(cm

)5.4�

1.2

(25)a

2.2

b6.1�

1.1

(43)a

1.2

–1.6

bN

/AN

/AN

/A

Fin

excurs

ion

(max

angle

)

38.6�

12.5

(23)c

N/A

107.1�

29.2

(32)c

121.1�

11.9

(7)d

N/A

P=

0.3

57

Po

0.0

001

Cycle

dura

tion

(s)

0.2

51�

0.0

25

(n=

23)c

0.2

35�

0.0

51

(n=

26)d

0.2

29�

0.0

49

(n=

32)c

0.0

2�

0.0

02

(n=

17)d

P=

0.0

18

Po

0.0

01

P=

0.0

13

Clim

bin

gbout

dura

tion

(s)

4.7�

3.1

(25)a

7.5

5�

6.4

5(1

7)e

2.7�

2.6

(43)a

1.9

7�

0.6

8(1

7)e

P=

0.0

16

P=

0.1

94

P=

0.0

01

Net

clim

bin

gspeed

(accounting

for

rest)

cm

s�

10.7

1�

0.4

9(2

5)a

0.2

2�

0.0

7(1

7)e

1.0�

1.3

(43)

0.8

2�

0.3

5(1

7)e

Po

0.0

01

P=

0.1

21

P=

0.0

87

body

length

s�

10.1

3�

0.0

88

(25)a

0.1

f0.0

68�

0.0

53

(43)a

0.5

85

fN

/AN

/APo

0.0

01

Clim

bin

g-o

nly

velo

city

(body

length

s�

1)

0.1

9�

0.1

2(2

5)a

0.2

1�

0.0

5(2

6)d

0.2

2�

0.2

1(4

3)a

12.4�

1.0

(17)d

P=

0.0

14

Po

0.0

01

P=

0.3

27

%tim

ein

motion

67�

16

(25)a

54�

21

(16)e

37�

19

(43)a

22�

11

(17)e

P=

0.1

24

P=

0.0

03

Po

0.0

01

All

sta

tisticalcom

parisons

perf

orm

ed

usin

gM

ann–W

hitney

Ute

sts

.S

am

ple

siz

es

are

report

ed

inpare

nth

eses

belo

weach

mean.

aD

ata

derived

from

30

Hz

DV

record

ings

inth

epre

sent

stu

dy.

bB

ody

length

sre

port

ed

by

Schoenfu

ss

&B

lob

(2003).

cD

ata

derived

from

hig

h-s

peed

vid

eo

record

ings

inth

epre

sent

stu

dy.

dD

ata

derived

from

hig

h-s

peed

vid

eo

record

ings

report

ed

by

Schoenfu

ss

&B

lob

(2003).

eD

ata

derived

from

30

Hz

DV

record

ings

report

ed

by

Blo

bet

al.

(2006).

f Valu

es

calc

ula

ted

based

on

mean

body

length

(Ss

=2.2

cm

,Lc

=1.4

cm

)re

port

ed

by

Schoenfu

ss

&B

lob

(2003)and

clim

bin

gspeed

(in

cm

s�

1)re

port

ed

by

Blo

bet

al.

(2006).


Ontogeny of waterfall climbing in gobiid fishesR. W. Blob et al.

JZO 315(BW

UK

JZ

O 3

15.P

DF

05-M

ar-0

7 19

:57

9547

12 B

ytes

10

PAG

ES

n op

erat

or=

Rav

isha

nkar

)

UNCORRECTED PROOF

(grains mean� SD=392� 99 mm, n=50) glued to the pipe.

Stream water was run over chutes for several days before

experiments to remove chemical residue, and a distance

scale was marked on each chute. To generate flow over the

climbing surface during trials, a siphon directed Hakalau

stream water down the chute from a 20L bucket at

200mLmin�1 (within natural flow rates: H. L. Schoenfuss,

pers. obs.). We began each trial by introducing three to five

fish (from a single species) into the basin with water flowing

down the chute. If climbing did not commence within

20min, the fish were removed and replaced with a new

group.

Two types of video records were collected. First, to

measure the climbing performance over a distance of several

body lengths, a SonyQ1 DTV 1020 digital camcorder (30Hz

framing rate) filmed animals in dorsal view as they ascended

the chute. Camera field of view covered 20 cm of pipe

c. 10 cm above the water level. Each fish was used in a single

trial and then returned to the stream. Trials were only

analyzed if the fish climbed the entire 20 cm. DV tape

footage was transferred to a computer (Macintosh G4)

using iMovie software. One of us (M. B.) tracked individual

animals frame by frame. The beginning and end of each

climbing bout were recorded, as were rest periods between

bouts and instances of pectoral fin and tail use during

climbing.

Second, to evaluate detailed climbing kinematics, a high-

speed digital video was collected for a subset of trials using

either a PhantomQ2 V4.1 (200Hz) or Redlake MotionscopeQ3 (250Hz) camera. Fish were filmed in dorsal view as they

climbed the chute and trials were saved as AVI files. The

positions of anatomical landmarks on the fishes were

digitized for each analyzed video frame using a modification

of the public domain NIH Image program for Macintosh,

developed at the US National Institutes of Health (the

modification, QuickImage, was developed by J. Walker and

is available at http://www.usm.maine.edu/�walker/software.html). For the inching climber S. stimpsoni, points at the

base and tip of both pectoral fins were digitized along with

four points on the head representing the maximal anterior,

left and right margins of the oral sucker, as well as the

midpoint between the eyes. At least nine additional points

along the body midline were also digitized, with the most

posterior point at the base of the caudal peduncle. For

species without an oral sucker, the same points were

digitized, except for those on head, for which the three

sucker landmarks were replaced with points at the anterior

midline of the face and the left and right eyes.

Custom programs (Matlab 5.0; Mathworks, Inc.; Natick,

MA, USA) calculated kinematic variables from digitized

coordinates. For all trials, fin movements and the displace-

ment of the front lip through each cycle were evaluated. The

angle of each fin to the direction of travel (determined by the

two midline head points) was calculated for each frame,

allowing us to calculate profiles of fin movements through-

out locomotor cycles and fin excursion angles. In addition,

for S. stimpsoni, an index of the instantaneous area of the

oral sucker was calculated as

Mouth area ¼ ðp=4ÞðDLRÞðDAPÞ ð1Þ

where DLR is the distance between the left and right edges of

the mouth and DAP is the distance from the point between

the eyes to the anterior edge of the mouth. This index

simplifies mouth shape, but allows evaluation of whether

the oral disk is increasing or decreasing in area (Schoenfuss

& Blob, 2003). After evaluating these parameters, we used

QuickSAND (Walker, 1998; available at http://www.usm.

maine.edu/�walker/software.html) to fit a quintic spline to

the kinematic calculations for each trial, smooth the data

and normalize all trials to the same duration to calculate the

mean kinematic profiles for each variable.

Statistical analyses were performed using GraphPad

Prism 4.0c (San Diego, CA, USA) or Statview 5.0 (SAS

Institute Inc, Cary, NC, USA) for Macintosh. The mean

locomotor cycle duration for each species was calculated

from a high-speed video, whereas the mean durations of

climbing and rest periods for entire climbing bouts (over the

20 cm distance) were calculated from standard video. Net

climbing speed (accounting for rest periods) was calculated

by dividing the 20 cm distance by the sum of climbing and

rest times. In addition, the percentage of time that each

individual spent moving was calculated as time in motion

divided by the sum of motion and rest times. Differences in

kinematic and performance variables between species and

between conspecific juveniles and adults were evaluated

using non-parametric Mann–Whitney U tests.

Results

Climbing by adult gobies: comparisons withjuveniles

Awaous guamensis

In contrast to powerbursting juveniles (Blob et al., 2006),

none of the 10 adult A. guamensis successfully climbed (or

Mini-DVcamera

High-speedcamera

Climbingchute

Catch basin

Siphon

Goby

Source bucketfor water

Figure 2 Illustration of experimental arena used for climbing trials with

adult Hawaiian stream gobies.



JZO 315(BW

UK

JZ

O 3

15.P

DF

05-M

ar-0

7 19

:57

9547

12 B

ytes

10

PAG

ES

n op

erat

or=

Rav

isha

nkar

)

UNCORRECTED PROOF

attempted to climb) the chute. As adults of the other two

focus species could climb, and adult A. guamensis behave

normally in captivity (R. W. Blob, T. Maie & H. L.

Schoenfuss, pers. obs.), the failure of A. guamensis to climb

is unlikely to be an artifact of our experimental design.

Sicyopterus stimpsoni

Much like juvenile S. stimpsoni, adults use inching move-

ments to climb waterfalls (Figs 3a–d and 4a–e), and begin

climbing cycles with a decrease in mouth area (Fig. 4b and d)

that reflects detachment of the oral sucker from the sub-

strate (Schoenfuss & Blob, 2003). During this time, there is

little forward advancement of the body in adults until

minimum mouth area is achieved (Fig. 4c and d). However,

unlike juveniles, in adult S. stimpsoni, each climbing cycle is

accompanied by pectoral fin excursions (Figs 3a–d and 4e).

From an initial position with tips pointed posteriorly, by

mid-cycle the pectoral fins abduct anteriorly to a nearly 401

angle from a vector pointing posteriorly along the body

(Table 1; Figs 3a–c and 4e). This position is held through the

middle third of the climbing cycle (Fig. 4e), during which

time the body begins to advance (Fig. 4c). Thus, initial

pectoral fin movements may put the fins in a position that

helps to hold adult S. stimpsoni in place on the substrate, but

(a) (b) (c) (d)

(h)(g)(f)(e)

Figure 3 Still images from a high-speed video of adult Hawaiian stream gobies climbing up the chute of the artificial waterfall arena. (a–d)

Sicyopterus stimpsoni. (e–h) Lentipes concolor. In each panel, the left pectoral fin of the fish is outlined with a dashed line, and time (in

milliseconds, ms) through a single climbing cycle for each species is indicated in the lower left corner of the panel.



JZO 315(BW

UK

JZ

O 3

15.P

DF

05-M

ar-0

7 19

:57

9547

12 B

ytes

10

PAG

ES

n op

erat

or=

Rav

isha

nkar

)

UNCORRECTED PROOF

initial forward movement appears to occur strictly through

advancement of the head. Advancement of adults through

the final third of the climbing cycle, however, is synchronous

with adduction of the pectoral fins back to their starting

position, suggesting that the pectoral fins might contribute

to climbing propulsion in adult S. stimpsoni in ways not

observed in juveniles. Some individuals made moderate

bending movements of the distal tail during some climbing

cycles, but this occurred irregularly. Because climbing pro-

ceeded for many cycles without these movements, they do

not appear to be necessary for successful climbing.

With regard to climbing performance, adults showed a

small (0.016 s), significant increase in climbing cycle dura-

tion compared with juveniles, but a much larger decrease

(�37%) in the duration of climbing bouts (i.e. sequences of

consecutive climbing cycles: Table 1). In absolute terms,

adult S. stimpsoni climbed the 20-cm distance significantly

(43 times) faster than juveniles; however, the adults eval-

uated were almost 2.5 times larger than juveniles, and this

difference in speed disappeared when speeds were normal-

ized by body length (Table 1). In fact, when climbing speed

was calculated based only on the duration of time that fish

were actually moving (i.e. without considering the time

spent resting between bouts of climbing), juveniles were

slightly (but significantly) faster than adults (Table 1).

Adults spent a greater portion of time in motion than

juveniles (67� 16 vs. 54� 21%), but high variance de-

creased our power to reject the null hypothesis of no

difference between adults and juveniles in this parameter.

Lentipes concolor

Lentipes concolor adults showed such dramatic changes in

climbing kinematics compared with juveniles that it is

inappropriate to describe adult L. concolor as ‘powerburst’

climbers. First, unlike juveniles, in adult L. concolor rapid

body undulations are absent during climbing (Fig. 3e–h).

Moreover, unlike juveniles in which the pectoral fins are

only adducted once to initiate a climbing bout (Schoenfuss

(a) (b)

(d) (e)

(g)(f)

(c)

Front lip displacement Mouth area Fin Excursion Angle

Sicyopterus stimpsoni juveniles

Sicyopterus stimpsoni adults

Lentipes concolor adults

Figure 4 Mean profiles of kinematic variables for Sicyopterus stimpsoni juveniles (a–b, n=26: data replotted for comparison with adults from

Schoenfuss & Blob, 2003), S. stimpsoni adults (c–e, n=23) and Lentipes concolor adults (f–g, n=32). Plots for the same variable are organized in

columns to facilitate comparison among these groups. Variables include cumulative displacement of the front lip since the start of the cycle,

measured in body lengths (a, c, f); instantaneous area of the mouth (i.e. oral sucker: b, d); and excursion angle of the pectoral fin (e, g). Each

climbing cycle was normalized to the same duration and values of kinematic variables were interpolated for 25 equally spaced increments through

the cycle using QuickSAND software (Walker, 1998). Plots for each variable illustrate means� 1 SE for each 4% increment of time through the

cycle. Fin excursion angle was not plotted for juvenile S. stimpsoni and mouth area was not plotted for adult L. concolor because these groups did

not produce the movements represented by these variables. No plots are shown for adult Awaous guamensis because climbing was not elicited

from adults of this species.



JZO 315(BW

UK

JZ

O 3

15.P

DF

05-M

ar-0

7 19

:57

9547

12 B

ytes

10

PAG

ES

n op

erat

or=

Rav

isha

nkar

)

UNCORRECTED PROOF

&Blob, 2003), in adults a cycle of pectoral fin abduction and

adduction accompanies every climbing cycle. Abduction

begins early in the cycle, advancing to 107.1� 29.21 from a

vector pointing posteriorly along the body, similar to

powerbursting juveniles (Figs 3f and 4g; Table 1). Pectoral

fin adduction begins soon after peak abduction is achieved

(Fig. 4g) and is closely followed by upward advancement of

the body at �60% through the climbing cycle (Figs 3g–h

and 4f). As in S. stimpsoni, some L. concolormade moderate

bending movements of the tail in some climbing cycles, but

this was irregular and climbing frequently took place with-

out lateral tail movements.

The climbing performance of adult L. concolor also

differs substantially from that of juveniles. Although the

durations of adult climbing cycles are significantly (410

times) longer than those of juveniles, the durations of

climbing bouts do not increase significantly among large

individuals (Table 1). Moreover, the absolute climbing

speed of adults (not normalized for body size) is also not

significantly different from that of juveniles (Table 1). As a

result, climbing speeds normalized for body size are more

than eight times slower in adult L. concolor than in juveniles

(Table 1). To achieve even this slower performance, adults

must spend a significantly greater portion of time moving,

rather than resting (37� 19%), compared with juveniles

(22� 11%; Table 1).

Interspecific comparisons of adult climbing

Although the two Hawaiian goby species that retain climb-

ing ability as adults (S. stimpsoni and L. concolor) used

different climbing styles as juveniles (inching vs. power-

bursts), as adults their kinematics and performance become

more similar. Pectoral fin excursion angles are significantly

greater in L. concolor (Table 1), but similarity is demon-

strated between the species because, unlike juveniles,

S. stimpsoni and L. concolor adults both incorporate pector-

al fin abduction/adduction sequences with every locomotor

cycle (Fig. 4). Similarly, although the mean cycle durations

are statistically different between adults of these species, the

difference between adult S. stimpsoni and L. concolor (0.251

vs. 0.229 s) is much less than that between their juveniles

(0.235 vs. 0.02 s; Table 1). Climbing bout durations show a

similar pattern, with the mean values for adults being

significantly different between species, but converging more

closely (4.7 vs. 2.7 s for S. stimpsoni and L. concolor,

respectively) than in juveniles (7.6 vs. 2.0 s).

Climbing speeds also become more similar in adults, with

L. concolor converging on S. stimpsoni (Table 1; Fig. 5a and

b). Lentipes concolor climb faster than S. stimpsoni as

juveniles (Blob et al., 2006), but adult S. stimpsoni speed

differs little from that of juveniles when normalized for body

length, whereas the speeds of adult L. concolor decrease

substantially from those of juveniles. As a result, adult

L. concolor are significantly slower than adult S. stimpsoni

(Table 1; Fig. 5b). However, one comparison that remains

similar in adults to patterns observed in juveniles is that

adult S. stimpsoni still spend a significantly longer portion of

time in motion than adult L. concolor (Table 1; Fig. 5c).

Discussion

Body size has pervasive effects on animal function, many of

which relate to changes in forces that animals experience as

1000.55

4

3

2

1

0

0.4

0.3

0.2

0.1

0.0

% T

ime

in m

otio

n

Vel

ocity

(B

I sec

−1)

Vel

ocity

(cm

sec

−1)

90

80

70

60

50

40

30

20

10

0

(a) (b) (c)

***

***

Figure 5 Box plots comparing (a) absolute climbing speed (in cm s�1), (b) relative climbing speed normalized for body length (in BL s�1) and (c)

portion of time spent moving (% time in motion) for adults of the Hawaiian stream gobies Sicyopterus stimpsoni (n=25) and Lentipes concolor

(n=43). Speeds are calculated over the entire 20-cm climbing distance (see text) and account for periods of rest between bouts. For each plot, the

box extends from the 25th to 75th percentile, with a line indicating the median. Bars demarcate the range of values. ��Significant difference at

Po0.001 (Mann–Whitney U-test).



JZO 315(BW

UK

JZ

O 3

15.P

DF

05-M

ar-0

7 19

:57

9547

12 B

ytes

10

PAG

ES

n op

erat

or=

Rav

isha

nkar

)

UNCORRECTED PROOF

they grow (Schmidt-Nielsen, 1984; McHenry & Lauder,

2005). Unless growing animals adopt mechanisms to accom-

modate these changing forces, the performance of many

functions might be impaired or precluded in larger indivi-

duals (Maie et al., in review). The three species of Hawaiian

gobies that climb waterfalls as juveniles show differing

functional responses as they increase in size. In A. guamen-

sis, climbing ability appears to have been lost in adults.

However, in the other two species of climbing Hawaiian

gobies, changes in locomotor kinematics preserve at least

some climbing ability among adults, allowing continued

opportunity for this novel behavior to meet the functional

requirements of a habitat with extreme physical demands.

Ontogenetic changes in waterfall climbing:mechanisms for accommodating larger bodysize

Three general mechanisms might enable animals to accom-

modate size-related increases in locomotor forces that they

experience (or must produce) due to growth: allometric

increases in morphological features (e.g. fin breadth); allo-

metric changes in physiological parameters (e.g. muscle

force generation); or changes in locomotor behavior.

Although morphological and physiological scaling have

not been completely surveyed in climbing gobies, at least

one feature integral to climbing, the area of the pelvic

sucker, fails to exhibit allometric growth that could help

accommodate the greater gravitational forces that must be

resisted by larger individuals as they climb (Maie et al., in

review). This led us to predict that if adult gobies could

climb, they would be expected to show behavioral adjust-

ments relative to juveniles that would facilitate maintenance

of climbing performance. Such adjustments appeared in

both species that continued to climb as adults. In adult S.

stimpsoni and L. concolor, pectoral fin movements were

added to every climbing cycle, contrasting with juveniles in

which the pectoral fins were used once (L. concolor) or not at

all (S. stimpsoni) during climbing bouts. In S. stimpsoni,

pectoral fin movements are less extensive than in L. con-

color, and fin adduction is delayed until after forward

advancement has commenced (Fig. 4c and e). Thus, exten-

sion of the fins may initially stabilize the body of adult

S. stimpsoni against the substrate during extension of the

head, only aiding forward propulsion toward the end of

each cycle. In contrast, the sweeping pectoral adductions of

adult L. concolor synchronize closely with body advance-

ment, and appear to provide the primary upward thrust in

this species in which axial motions are reduced compared

with powerbursting juveniles.

Size-related changes in climbing performance and kine-

matics among species using powerbursts as juveniles were

more drastic than those in the inching climber S. stimpsoni:

one powerburst species (A. guamensis) did not climb as

adults, and the other (L. concolor) shifted from using the

body axis as its main propulsor to primary reliance on the

pectoral fins. This divergence in climbing performance and

mechanics between juveniles and adults suggests that power-

bursts may not be a viable climbing mechanism for larger

individuals. Adult A. guamensis and L. concolor still appear

capable of performing body movements associated with

powerburst climbing: body axis movements seen in juvenile

climbers (Schoenfuss & Blob, 2003) are seen in swimming

adults (R. W. Blob, H. L. Schoenfuss, pers. obs.), and

pectoral fin movements used by climbing adult L. concolor

simply represent more frequent use of a behavior exhibited

by juveniles (Table 1). However, adult powerburst climbing

might be precluded because of its high demand for muscular

power (Blob et al., 2006). Ascension of vertical surfaces has

high power requirements (Irschick et al., 2003), but muscles

contracting at the high frequencies used during powerburst

climbing tend to have reduced power-producing capacities

(Rome, 1998). Because mass-specific power production also

tends to decrease with body size (Irschick et al., 2003), it

might not be possible to climb using rapid movements at

larger body sizes, leaving slower movements like pectoral fin

crutching and the inching of S. stimpsoni as the only viable

climbing mechanisms for adult gobies. Such size-related

limitations could also help explain the lack of powerburst

climbing in juvenile S. stimpsoni, which can be three times

heavier than juveniles from powerburst species and possibly

already restricted in climbing behaviors (Schoenfuss, 1997).

Physiological implications of climbingperformance in adult Hawaiian streamgobies

Like juveniles, both species that successfully climb waterfalls

as adults use intermittent bouts of locomotion. For other

species moving over terrestrial substrates, the use of inter-

mittent locomotion significantly increases the distance tra-

veled before fatigue (Weinstein & Full, 1998, 1999). This

benefit results if pauses are of appropriate duration (abso-

lute and relative to time spent moving) to allow removal of

fatigue-inducing products and restoration of metabolic fuels

(Allen & Westerblad, 2001; Kramer & McLoughlin, 2001).

Intermittent inching by S. stimpsoni juveniles did not meet

these criteria (Blob et al., 2006), as activity durations ranged

up to twice as long as rest periods, a ratio that does not

typically improve performance (Weinstein & Full, 1998,

1999). Adult S. stimpsoni did not differ significantly from

juveniles in the ratio of time spent climbing versus resting

(Table 1); thus, they also seem unlikely to increase the

distance traveled before fatigue through intermittent climb-

ing. In contrast, juvenile powerburst climbers rest between

three and five times longer than they move (Blob et al.,

2006). With the short durations of their activity periods

(o2 s), the motion-to-rest ratio of juvenile powerburst

climbers is appropriate to increase their pre-fatigue travel

distance (Weinstein & Full, 1998, 1999) at the cold tempera-

tures typical of their stream habitats (18–19 1C: Schoenfuss,

Julius & Blob, 2004). Although adult L. concolor rested for a

significantly shorter portion of time than juveniles, they still

rested for over 1.7 times longer than their activity periods

(i.e. motion-to-rest ratio=0.59); in addition, adult locomo-

tor bouts did not last significantly longer than those of



JZO 315(BW

UK

JZ

O 3

15.P

DF

05-M

ar-0

7 19

:57

9547

12 B

ytes

10

PAG

ES

n op

erat

or=

Rav

isha

nkar

)

UNCORRECTED PROOF

juveniles (Table 1). This ratio of activity to rest is close to

0.5, which the increases distance capacity in species ranging

from crabs to geckos (Weinstein & Full, 1998, 1999). Thus,

although adult L. concolor use climbing mechanisms differ-

ent from juveniles, they may still achieve some of the same

performance benefits.

Slow climbing by adult gobies may also help prevent

fatigue. Climbing cycle frequencies for adult S. stimpsoni

and L. concolor (calculated from cycle durations, Table 1)

average 4.0 and 4.4Hz, respectively, and climbing speeds

(considering motion periods only) are o0.25BL s�1 (Table

1). Motion at such low speeds and frequencies in fish is

typically powered by slow oxidative (red) muscle without

activity in other fiber types (Jayne & Lauder, 1994; Gillis,

1998). If adult gobies used only red muscle during climbing,

their potential for fatigue could be reduced because red

muscle activity depends on aerobic metabolism that is suited

for endurance behaviors (Sanger & Stoiber, 2001). Extensive

fast glycolytic (white) fibers would still be expected in adults,

however, as these would facilitate swimming in rapid cur-

rents, likely their most common locomotor behavior.

Ecological implications of climbingperformance in adult Hawaiian streamgobies

The climbing performance exhibited by adult stream gobies

may relate to several aspects of their ecology. First, because

A. guamensis has the poorest climbing juveniles among the

three climbing species (Blob et al., 2006), it is not surprising

that adults of this species have lost climbing ability that was

retained by the other species. Moreover, the loss of climbing

among larger A. guamensis would help explain why this

species showed the greatest delay in the upstream appear-

ance of reproductive adults in streams on Kaua’i that were

devastated by Hurricane Iniki in 1992 (Fitzsimons & Nishi-

moto, 1995), as repopulation may have been achieved solely

by small individuals and juveniles migrating in from the

ocean.

One of the most striking features of the climbing exhib-

ited by juvenile Hawaiian gobies is the diversity of mechan-

isms and levels of performance that they use. Not only do

juveniles use two dramatically different climbing styles

(inching vs. powerbursts), but different species of power-

burst climbers exhibit different performance. This diversity

resembles the ‘many-to-one’ mapping of morphology to

function documented in fish feeding studies (Alfaro, Bolnick

&Wainwright, 2005; Wainwright et al., 2005), and indicates

that a wide range of performance can be maintained even

under the selective pressures of extreme environmental

conditions (Blob et al., 2006). However, functional diversity

is diminished among adult climbers, suggesting a ‘fewer-to-

one’ mapping of performance to functional demand.

Although there are statistically significant differences in

several performance variables between adult S. stimpsoni

and L. concolor, the magnitude of those differences is

smaller between adults than between juveniles (Table 1).

Moreover, adults of both species incorporate the same

additional propulsive mechanism (pectoral fin adduction)

beyond those seen in juveniles (Table 1). Thus, adult

S. stimpsoni and L. concolor have converged considerably

in their strategies and capacity for climbing waterfalls.

Although functional diversity might be maintained among

juveniles of these species despite strong selection for success-

ful waterfall climbing, the additional pressures of increasing

body size may finally limit the range of mechanics and

performance that can execute this novel behavior.

Acknowledgements

Funding was provided by the Hawai’i DAR (Sport Fish

Restoration Project F-14-R-28) to H.L.S., M.L.J. and

R.W.B.; the Clemson Department of Biological Sciences; a

Raney Award from the American Society of Ichthyologists

and Herpetologists (T.M.); a St Cloud State Faculty Re-

search Grant (H.L.S.); and Stearns Manufacturing Inc. We

thank Bob Nishimoto, Darrell Kuamo’o, Mike Fitzsimons

and Nora Espinoza for suggestions, Christopher Mims for

data analysis and Mike Yamamoto for the photos used in

Fig. 1b and d. We also thank Lance Nishiura, Tim Shindo,

Troy Sakihara, Troy Shimoda and Bruce Kaya for logistical

support at the Hilo DAR station.

References

Alfaro, M.E., Bolnick, D.I. & Wainwright, P.C. (2005).

Evolutionary consequences of many-to-one mapping of

jaw morphology to mechanics in labrid fishes. Am. Nat.

165, E140–E154.

Allen, D.G. & Westerblad, H. (2001). Role of phosphate and

calcium stores in muscle fatigue. J. Physiol. 536, 657–665.

Blob, R.W., Rai, R., Julius, M.L. & Schoenfuss, H.L. (2006).

Functional diversity in extreme environments: effects of

locomotor style and substrate texture on the waterfall

climbing performance of Hawaiian gobiid fishes. J. Zool.

(Lond.) 268, 315–324.

Carrier, D.R. (1996). Ontogenetic limits on locomotor per-

formance. Physiol. Zool. 69, 467–488.

Ego, K. (1956). Life history of fresh water gobies. Project No.

4-4-R. Fresh water game fish management research, Hon-

olulu, Territory of Hawai’i.

Fitzsimons, J.M. & Nishimoto, R.T. (1995). Use of fish

behavior in assessing the effects of Hurricane Iniki on the

Hawaiian island of Kaua’i. Environ. Biol. Fish. 43, 39–50.

Gillis, G.B. (1998). Neuromuscular control of anguilliform

locomotion: patterns of red and white muscle activity

during swimming in the American eel, Anguilla rostrata.

J. Exp. Biol. 201, 3245–3256.

Higham, T.E., Malas, B., Jayne, B.C. & Lauder, G.V. (2005).

Constraints on starting and stopping: behavior compen-

sates for reduced pectoral fin area during breaking of the

bluegill sunfish Lepomis macrochirus. J. Exp. Biol. 208,

4735–4746.



JZO 315(BW

UK

JZ

O 3

15.P

DF

05-M

ar-0

7 19

:57

9547

12 B

ytes

10

PAG

ES

n op

erat

or=

Rav

isha

nkar

)

UNCORRECTED PROOF

Irschick, D.J., Vanhooydonck, B., Herrel, A. & Andronescu,

A. (2003). Effects of loading and size on maximum power

output and gait characteristics in geckos. J. Exp. Biol. 206,

3923–3934.

Jayne, B.C. & Lauder, G.V. (1994). How fish use slow and fast

muscle fibers: implications for models of vertebrate muscle

recruitment. J. Comp. Physiol. A 175, 123–131.

Julius, M.L., Blob, R.W. & Schoenfuss, H.L. (2005). The

survival of Sicyopterus stimpsoni, an endemic amphidro-

mous Hawaiian gobiid fish, relies on the hydrological

cycles of streams: evidence from changes in algal composi-

tion of diet through growth stages of fish. Aquat. Ecol. 39,

473–484.

Kramer, D.L. & McLoughlin, R.L. (2001). The behavioral

ecology of intermittent locomotion. Am. Zool. 41,

137–153.

Maie, T., Schoenfuss, H.L. & Blob, R.W. (In review). Onto-

genetic scaling of body proportions in waterfall-climbing

gobiid fishes from Hawai’i and Dominica: implications for

locomotor functionQ4 .

McHenry, M.J. & Lauder, G.V. (2005). Ontogeny of form

and function: locomotor morphology and drag in zebrafish

(Danio rerio). J. Morphol. 267, 1099–1109.

Nishimoto, R.T. & Kuamo’o, D.G.K. (1992). Waterfall

climbing behavior in the postlarvae of native Hawaiian

stream gobies. Technical paper, Division of Aquatic

Resources, State of Hawai’i.

Rome, L.C. (1998). Some advances in integrative muscle

physiology. Comp. Biochem. Physiol. B 120, 51–72.

Sanger, A.M. & Stoiber, W. (2001). Muscle fiber diversity and

plasticity. In Fish physiology, Vol. 18: 187–250. Johnston,

I.A. (Ed.). San Diego: Academic Press.

Schmidt-Nielsen, K. (1984). Scaling: why is animal size so

important? Cambridge: Cambridge University Press.

Schoenfuss, H.L. (1997). Metamorphosis of the Hawaiian

stream goby Sicyopterus stimpsoni: a structural, functional,

and behavioral analysis. PhD thesis, Louisiana State Uni-

versity.

Schoenfuss, H.L. & Blob, R.W. (2003). Kinematics of water-

fall climbing in Hawaiian freshwater fishes (Gobiidae):

vertical propulsion at the aquatic–terrestrial interface.

J. Zool. (Lond.) 261, 191–205.

Schoenfuss, H.L., Julius, M.L. & Blob, R.W. (2004). Coloni-

zation of a recent, volcanically formed freshwater habitat:

an example of primary succession. Ichthyol. Explor. Fresh-

waters 15, 83–90.

Toro, E., Herrel, A., Vanhooydonck, B. & Irschick, D.J.

(2003). A biomechanical analysis of intra- and interspecific

scaling of jumping and morphology in Caribbean Anolis

lizards. J. Exp. Biol. 206, 2641–2652.

Wainwright, P.C., Alfaro,M.E., Bolnick, D.I. &Hulsey, C.D.

(2005). Many-to-one mapping of form to function: a

general principle in organismal design? Int. Comp. Biol. 45,

256–262.

Walker, J.A. (1998). Estimating velocities and accelerations of

animal locomotion: a simulation experiment comparing

numerically different alogorithms. J. Exp. Biol. 201, 981–995.

Weinstein, R.B. & Full, R.J. (1998). Performance limits of

low-temperature, continuous locomotion are exceeded

when locomotion is intermittent in the ghost crab. Physiol.

Zool. 71, 274–284.

Weinstein, R.B. & Full, R.J. (1999). Intermittent locomotion

increases endurance in a gecko. Physiol. Biochem. Zool. 72,

732–739.



JZO 315(BW

UK

JZ

O 3

15.P

DF

05-M

ar-0

7 19

:57

9547

12 B

ytes

10

PAG

ES

n op

erat

or=

Rav

isha

nkar

)

Author Query Form

_______________________________________________________

_______________________________________________________

Dear Author,

During the copy-editing of your paper, the following queries arose. Please respond to these by marking up your proofs with the necessary changes/additions. Please write your answers clearly on the query sheet if there is insufficient space on the page proofs. If returning the proof by fax do not write too close to the paper's edge. Please remember that illegible mark-ups may delay publication.

Journal JZOArticle 315

Query No. Description Author Response

Q1AQ: Please provide manufacturer information for Sony: town, state (if USA) and country.

Q2AQ: Please provide manufacturer information for Phantom: town, state (if USA) and country.

Q3AQ: Please provide manufacturer information for Redlake Motionscope: company, town, state (if USA) and country.

Q4AQ: Please update the reference Maie et al. (in review) if accepted for publication or remove from the list if not accepted.

MARKED PROOF

Please correct and return this set

Instruction to printer

Leave unchanged under matter to remain

through single character, rule or underline

New matter followed byor

or

or

or

or

or

or

or

or

and/or

and/or

e.g.

e.g.

under character

over character

new character new characters

through all characters to be deleted

through letter orthrough characters

under matter to be changedunder matter to be changedunder matter to be changedunder matter to be changedunder matter to be changed

Encircle matter to be changed

(As above)

(As above)

(As above)

(As above)

(As above)

(As above)

(As above)

(As above)

linking characters

through character orwhere required

between characters orwords affected

through character orwhere required

or

indicated in the marginDelete

Substitute character orsubstitute part of one ormore word(s)

Change to italicsChange to capitalsChange to small capitalsChange to bold typeChange to bold italicChange to lower case

Change italic to upright type

Change bold to non-bold type

Insert ‘superior’ character

Insert ‘inferior’ character

Insert full stop

Insert comma

Insert single quotation marks

Insert double quotation marks

Insert hyphenStart new paragraph

No new paragraph

Transpose

Close up

Insert or substitute spacebetween characters or words

Reduce space betweencharacters or words

Insert in text the matter

Textual mark Marginal mark

Please use the proof correction marks shown below for all alterations and corrections. If you

in dark ink and are made well within the page margins.wish to return your proof by fax you should ensure that all amendments are written clearly

Documents

Ontogenetic change in novel functions: waterfall climbing