Pacific Science (1978), vol. 32, no. 2 © 1978 by The University Press of Hawaii. All rights reserved

Growth and Size of the Tropical (Halodeima) atra Jager at Enewetak Atoll, Marshall Islands!

THOMAS A. EBERT 2

ABSTRACT: In September 1975, 5031 sea cucumbers (Holothuria atra Jager) were tagged with tetracycline and returned to a gutter on the seaward reef bench of Ananij Island, Enewetak Atoll, Marshall Islands. A sample of 184 individuals was collected in September 1976. were dissected and plates of the calcareous ring were examined with ultraviolet radiation for tetracycline lines. Based on 18 tagged individuals, the Brody-Bertalanffy growth constants are: K = 0.11 and P00 = 0.89 cm (for interradial plates). Length (L, cm) and weight (W, g) of individuals are related to plate size: L = 36.35P and W = 1950P3, giving maximum size as 32 cm and 1352 g. Length frequency distributions for the population did not change from 1975 to 1976. A preliminary estimate of annual loss is 50 to 70 percent of the total population based on the growth parameters and population size structure. Holothuria atra shows asexual reproduction by transverse fission, which appears to be the major source of recruitment at Ananij rather than from the plankton.

ON MANY TROPICAL REEFS, sea cucumbers are tion to a technique that shows some promise abundant and may attain densities in excess as a means ofstudying growth ofholothurians of 200/m2 (Bakus 1968). But they are very and so, somewhat indirectly, of estimating difficult to study as populations and very few mortality and population turnover. The estimates of growth and death rates of holo- experimental was Holothuria (Halo­ thurians are in the literature. Edwards (1908) deima) atra Jager, an abundant and widely presents growth data for newly metamor- distributed Indo-Pacific aspidochirote, which phosed Holothuria floridana Pourtales, but was studied at Ananij Island, Enewetak Atoll only up to a size of 4 mm and an age of 75 (11°28' N, 162°24' E). A second purpose of days under laboratory conditions. Ruther- this paper is to present information about ford (1973) presents growth data of newly growth, death, and size for this species. recruited Cucumaria pseudocurata to an age Holothuria atra is very abundant at certain of 1 year and gives estimates of loss rate for areas ofEnewetak Atoll. Bakus (1973) shows animals less than I year old. Fish (1967) and a picture with a density of 5 to 35 individuals Buchanan (1967) discuss growth and mor- 1m 2 at what he calls Anayaanii Island. This tality of Cucumaria elongata and calculate is the same island Johannes et al. (1972) call rates based on data for animals of many Japtan and which is now called Ananij. ages. Military charts list this island as Bruce. The ___--"T~h=e_I2UrI2_QSLQLthis-papeLis-to-cal-l-att~n-aFl'}a-w-hefe-eueumbers-were-studtetlseems to be the same area illustrated by Bakus (1973); it is a reef gutter on the seaward side of

1 This work was supported by an Energy Research Ananij. The vegetated portion of Ananij is and Development Administration grant to the Univer­ narrow at the southern end and has an sity of Hawaii and National Science Foundation grant enclosed inlet on the seaward side bordered DES75-10442 to the author. Manuscript accepted 2 February 1978. by the main part of the island on the north 2 San Diego State University, Department of Biology, and by a narrow peninsula of vegetated land San Diego, California 92182. (covered nearly exclusively by Pemphis aci- 183 184 PACIFIC SCIENCE, Volume 32, April 1978 dula) on the south. The gutter where sea to remove the calcareous ring. An additional cucumbers were tagged is to the south of this 68 animals were measured and dissected in vegetated peninsula. At low tide the gutter is the field, so measurements would be on separated from the sea, but at high tide there animals handled the same as those used to is a strong current flowing from east to west determine size structure of t.he population. and deflected to the south and around the . Rings that were dissected from H. atra southern tip of the island. The bottom of the were treated with 5 percent sodium hypo­ gutter is an eroded reef with a layer of 0 to 5 chlorite to remove organic material, washed, cm of coral sand. driea, and then examined with ultraviolet radiation under a dissecting microscope.

METHODS RESULTS Tetracycline is known to be taken up by the growing surfaces ofechinoderm skeletons Of the 184 animals collected in 1977, only (Kobayshi and Taki 1969, Pearse and Pearse 18 individuals showed clear tetracycline 1975, Taki 1971) and has been used to analyze marks. I have reason to believe, however, growth in echinoids (Ebert 1977). Holo- that all animals I dissected had been tagged. thurians have, in addition to small ossicles Sea cucumbers were very highly concentrated in the body wall, a series oflarger plates at the in the areas where they had been released in beginning of the pharynx called the cal- 1975, but densities were very low in the careous ring (Hyman 1955). The plan was to surrounding areas where individuals had tag these plates of the calcareous ring with been collected. The recovery of only 10 tetracycline, measure growth of these plates, percent with a clear mark, I believe, is due and use this information to describe growth to only 10 percent of the population actually of other body parts of H. atra. . growing at the time of tagging or to regener- On 20 and 21 September 1975, 5031 ation rather than mixing of tagged and individuals were tagged with 0.2 ml of a untagged sea cucumbers. Tetracycline hydro­ solution of 1 g tetracycline hydrochloride chloride is both temperature- and pH­ dissolved in 100 ml seawater. The solution sensitive; the higher the pH and temperature, was injected into the body with no attempt the faster breakdown occurs (Barnes 1971). to select a definite site. Length of the first Breakdown on a tropical reef probably takes 1015 sea cucumbers was measured to the no longer than 1 day, so if an animal is not nearest 0.5 em. Holothurians are able to growing at the time it is tagged, it will not be change length and weight, so the only hope marked. Also, animals that regenerated a of obtaining reasonable comparisons of size complete ring after tagging would lack a was to try to standardize the handling tech- mark. nique. Animals were collected and placed in a Lengths of the interradial pieces of the plastic laundry basket and when the basket calcareous ring were measured using an was full (about 200 animals) it was moved ocular micrometer [see Hyman (1955), Figure about 15 meters to the place where the 56, for an illustration of plates of the cal­ animals were measured as quickly as possible. careous ring]. Length of the plate is defined Sea cucuml:5ers were pllrc"eu-on-a--f1sh---as-t-he-Elista-nGe--a-lQug-the-ramuLoLtb.e ring measuring board and length was read within (i.e., perpendicular to the long axis of the 10 sec. There appeared to be no difference in individual). Two to four plates were measured tendency to contract or relax between animals in each individual. In a sample of 30 animals at the top or bottom of the basket. with average plate lengths of 0.426 em and On 22-28 September 1976, 610 animals SD = 0.059, the average standard error of were measured in the field; 116 were brought plate length within individuals was 0.024 cm, back to the marine laboratory on Enewetak with an associated standard deviation of Island where they were weighed and dissected 0.015 em. It is apparent that variation of Tropical Sea Cucumber at Enewetak-EBERT 185

5 -E -E I I- 4 (!) Z W -.J W 3 «I- -.J Q. «-.J 2 z u.

1

1 2 3 4 5 6 INITIAL PLATE LENGTH (mm) FIGURE I. Ford-Walford plot of tetracycline-tagged interradial plates of the calcareous ring of Holothuria atra, Ananij Island, Enewetak. Plate length in 1976 (final plate length) is plotted against the length of the internal glowing image (initial plate length, i.e., the size in 1975). Dashed lines A and Care 95 percent confidence limits of slope. The line drawn at 45° is the line of zero growth. plate sizewithin an individual was substantial of the Brody-Bertalanffy growth equation __and_also.ihat-some-indi¥-iduals-were-mueh-less---fsee-;-foT-example-;--

24

20 • -E - -0 ·...- -.--:/i. ., .-z 16 ••• ., C) -;\":¢.-- ., Z w .- . ., -oJ • (it •• 12 • •., -oJ • • • ., .-

4

1 2 3 4 5

PLATE LENGTH (mm)

FIGURE 2. Total length of Holothuria atra individuals at Ananij Island versus length of interradial plates. Total length measurements were made in the field using the same handling procedure as was applied in constructing length frequency distributions. tion are calculated from the slope and inter­ length, P, measured in centimeters, and time, cept of the Ford-Walford line. Because the t, in years, is regression line rotates around the point (St, Pt = 0.891(1 - be-O.I06t) (1) t+ I )0neltne-withLhe-largest-slope-als0-has the smallest intercept; in terms of Brody­ The parameter b is equal to (P 00 - Po)/P00, Bertalanffy constants, a large growth rate where Po is the size of a plate in a newly constant, K, will be associated with a small settled individual. I have assumed that Po = asymptotic size, P00. Using the regression 0.0, so b = 1.0. Actual size of a plate in such line (B) in Figure 1, parameter estimates of an individual is probably about 0.001 em, the growth equation are K = 0.106 and P 00 so b would be equal to 0.999. Assuming b = = 0.891 em. The growth equation for plate 1.0 does not affect positioning of the curve Tropical Sea Cucumber at Enewetak-EBERT 187

to a significant degree. For a slope of 0.728 for both years are unimodal and are not (the lower 95 percent limit) and intercept of skewed. The two distributions were compared 0.148 em, the parameters would be K = for differences in location of the mean or in 0.317 and P oo = 0.542. shape using the Kolomogrov-Smirnov test The parameter K will have a negative sign (Tate and Clelland 1957). The maximum and there is no asymptotic size when the difference between the two distributions was Ford-Walford line has a slope greater than 4.9 percent at the interval 11 to 12 em. The 1.0. The Brody-Bertalanffy equation may significant difference with a = 0.05 is 6.97 still be used, but should be rewritten without percent (n 1 = 1015; nz = 610), so the null using the symbol P 00 : hypothesis H o is not rejected and I conclude that the size distribution had not changed KI C(1 - be- ) from 1975 to 1976. PI = K (2) 1 - e Size distributions can be used in con­ junction with growth parameters to obtain where C is the y intercept and e- K is the initial estimates of mortality or loss. Van slope of the Ford-Walford line. The upper Sickle (1977) provides an easily applied 95 percent limit of the regression coefficient technique for estimating m, the instantaneous in Figure 1 is 1.058, so K = -0.056 and C = loss rate: 0.042. Finally, with a slope of 1.0, the growth equation is m = -gs + K (6) PI = 0.06t (3) where s is the slope of the right descending limb of the semi-loge transformed size dis­ Growth was expressed in terms of total tribution (In N versus size), g is the average body length by establishing the relationship growth rate of animals from the mode to the between body length and plate length. Figure largest indivduals, and K is the growth 2 plots length for the 68 individuals measured constant. in the field against plate length, and shows a Using combined size distributions from zero intercept regression line (Bliss 1967). Figure 4, the slope of In N versus size is Where L is total length in em, P is plate -0.549. For line B in Figure 3, the average length in em, and with the 95 percent con­ growth rate is 1.66 em/year for animals 13 fidence interval of the slope, the relationship to 22 em long. The value of K is 0.106; IS accordingly, m = 1.02. The average growth L = 36.35( ± l.30)P rates for lines A and Care 2.45 and 0.60 (4) em/year, which give instantaneous loss rates Plate length also was related to total wet of 1.29 and 0.65, respectively. Annual loss weight using 99 animals brought back to the rates are 72, 64, and 48 percent for lines A, marine laboratory at Enewetak Island. A B, and C, respectively, in Figure 3. regression line was fitted with zero intercept, Loss of individuals is only in part due to with weight, W, in grams, and with the 95 mortality. Holothuria atra appears to be percent confidence interval for the slope: capable of asexual reproduction by trans- W = 1950(± 178)P3 (5) verse fission. This has been reported in the literature (Bonham and Held 1963) and ___,Gwwth-Gur-v€-s-fer-tel:aHength-are-shown-reviewed-Dy-B"akus (T973)-for other holo- ­ in Figure 3. An average animal 5 years old thurians. I collected several small H. atra at would be between 8.5 and 15.5 em long and Ananij between 5.0 and 7.0 em long, all of weigh between 25 and 156 g; a 9-year-old which were very wide for their length and individual would be between 17.0 and 20.0 so fit the description given for small in­ em long and weigh between 200 and 320 g. dividua1s by Bonham and Held (1963). One Length frequency distributions for 1975 individual had very large plates in its cal­ and 1976 are shown in Figure 4. Distributions careous ring, one had exceptionally small 188 PACIFIC SCIENCE, Volume 32, April 1978

A/ / 30 / / / / /B / 25 / /

-E 20 -() I ~ (!Jz 15 w ....J

10

5

2 4 6 8 AGE (YEARS) FIGURE 3. Brody-Bertalanffy growth curves for Holothuria atra using constants obtained from Figure I and the conversions of interradial plate to total length. Line A is from A in Figure 1, line B is from B in Figure 1, etc. plates, and one had no ring at all. A reason- careous ring in sea cucumbers. The two major able interpretation is that the individual with problems with the technique are (1) the the large plates was the front end ofa divided necessity of sacrificing the animals in order individual; the other two were back ends in to determine whether they were tagged and different stages of regeneration. also to measure the amount of growth, and Asexual reproduction has important con- (2) an animal showing zero growth during a ___~uences when interpreting the loss rates. year cannot be distinguished from an in­ In part, loss oflarge ioolvloualrfrom-a---dividua1-t-hat-was-ta-ggea-anEl-n(}t-gn)w-ing_ population would not be through death but but grew at some other period during the by fission. year, or from an individual that regenerated its plate or never was tagged. As indicated by Bakus (1973), the existence of asexual reproduction by transverse fission DISCUSSION complicates the analysis of size distributions. Use of tetracycline shows some promise Sizes that are normally distributed are shown for tagging the growing pieces of the cal- by Bonham and Held (1963) for H. atra at Tropical Sea Cucumber at Enewetak-EBERT 189

15

10 20 SEPT. 1975 N = 1015

5 -ae -> 0 z 20 25 w ::> 0 w a: 15 LL

10 22 SEPT. 1976 N = 610

5

15

LENGTH (em)

FIGURE 4. Length frequency distributions for Hololhuria alra at Ananij Island for 1975 and 1976. These distri­ butions are not significantly different (0( = 0.05) using the Kolomogrov-Smimov two-sample test. 190 PACIFIC SCIENCE, Volume 32, April 1978

Kabelle Island, Rongelap Atoll. The mean ---. 1973. The biology and ecology of length at Kabelle was about 13.0 cm, with tropical holothurians. BioI. Ecol. Coral minimum and maximum values of 5.0 cm Reefs 2 (Bio. 1): 325-367. and 25.0 cm, respectively. It is a distribution BARNES, D. J. 1971. A study of growth, that looks much like the size structures of the structure and form in modern coral skele­ cucumbers at Ananij. Also, like H. atra at tons. Ph.D. Thesis. University ofNewcastle Enewetak, they found very large individuals upon Tyne. 180 pp. at other locations around the atoll. The BLISS, C. I. 1967. Statistics in biology. Vol. largest individual they report at Rongelap I. McGraw-Hill, New York. 558 pp. was 60 cm; the largest individual I have BONHAM, K., and E. E. HELD. 1963. Ecologi­ measured at Enewetak was in the quarry on cal observations on the sea cucumbers Enewetak Island and was 42 cm long. Holothuria atra and H. leucospilota at It is possible that large individuals are not Rongelap Atoll, Marshall Islands. Pac. found in certain areas because transverse Sci. 17:305-314. fission takes place at small and medium sizes. BUCHANAN, J. B. 1967. Dispersion and demo­ Certain environmental features could pro­ graphy ofsome infaunal popu­ mote asexual reproduction, and in such lations. Symp. Zool. Soc. London 20: I-II. locations only small to medium size animals EBERT, T. A. 1975. Growth and mortality would be found. of post-larval echinoids. Am. Zool. It is interesting that in the gutter at Ananij 15: 755-775. no new individuals appear to have come ---. 1977. An experimental analysis of from the plankton. The same appears to be sea urchin dynamics and community in­ true for Bonham and Held's animals at teractions on a rock jetty. J. Exp. Mar. Kabelle Island. All small individuals were BioI. Ecol. 27: 1-22. very wide and probably represent fission EDWARDS, C. L. 1908. Variation, develop­ products. A reasonable interpretation is that ment and growth in Holothuria floridana recruitment from the plankton is very un­ Pourtales and in Holothuria atra Jager. predictable and asexual reproduction is an Biometrika 6:236-301. adaptation which permits the species to FISH, J. D. 1967. The biology of Cucumaria persist on a reef and span the periods of time elongata (Echinodermata: Holothuroidea). during which no settlement from the plank­ J. Mar. BioI. Assoc. U.K. 47: 129-143. ton takes place. The apparent stability of the GULLAND, J. A. 1969. Manual ofmethods for size structure at Ananij coupled with the fish stock assessment. Part 1. Fish popula­ estimated high loss rates and lack of recruit­ tion analysis. FAO Man. Fish. Sci. 4. ment from the plankton can be reconciled 154 pp. only by concluding that the rate of fission HYMAN, L. H. 1955. The invertebrates. Vol. must be high. IV. Echinodermata. McGraw-Hill, New York. 763 pp. JOHANNES, R. E., J. ALBERTS, C. D'ELIA, ACKNOWLEDGMENTS R. A. KINZIE, L. R. POMEROY, W. SOTTILE, Assistance in the field was provided by J. W. WIEBE, J. A. MARSH, Jr., P. HELFRICH, Harrison, J. Mynderse, and T. Turk. The J. MARAGOS, J. MEYER, S. SMITH, D. manuscript benefited from the critical revlew--€R*BTREE,A~R:0"FH,1r.--R~MGCbQSKEY,- of D. M. Dexter and one anonomyous S. BETZER, N. MARSHALL, M. E. Q. PILSON, reviewer. G. TELEK, R. I. CLUTTER, W. D. DUPAUL, K. L. WEBB, and J. M. WELLS, JR. 1972. The metabolism of some coral reef com­ LITERATURE CITED munities: A team study of nutrient and energy flux at Enewetak. Biosci. 22: 541­ BAKUS, G. J. 1968. Defense mechanisms and 543. ecology of some tropical holothurians. KOBAYASHI, S., and J. TAKI. 1969. Calcifica­ Mar. BioI. 2: 23-32. tion in sea urchins. 1. A tetracycline in-

<. Tropical Sea Cucumber at Enewetak-EBERT 191

vestigation of growth of the mature test in Cucumaria pseudocurata. Mar. BioI. 22: Strongylocentrotus intermedius. Calc. Tiss. 167-176. Res. 4: 210-223. TAKI, J. 1971. Tetracycline labelling of test PEARSE, J. S., and V. B. PEARSE. 1975. Growth plates in Strongylocentrotus intermedius. zones in the echinoid skeleton. Am. Zool. Sci. Rep. Hokkaido Fish. Exp. Sta. 13: 19­ 15:731-753. 29. RICKER, W. E. 1975. Computation and inter­ TATE, M. W., and R. C. CLELLAND. 1957. pretation of biological statistics of fish Nonparametric and shortcut statistics. In­ populations. Fish. Res. Board Can. Bull. terstate Printers and Publishers, Danville, 191. 382 pp. Ill. 171 pp. RUTHERFORD, J. C. 1973. Reproduction, VAN SICKLE, J. 1977. Mortality rates from growth, and mortality of the holothurian size distributions. Oecologia 27:311-318.