The relationship between temperature, growth rate, and duration of planktonic life for larvae of the gastropod Crepidula fornicata (L.)

241

J. Exp. Mar. Biol. Ecol., 1984, Vol. 14, pp. 241-257

Elsevier JEM 204

THE RELATIONSHIP DURATION

BETWEEN TEMPERATURE,

OF PLANKTONIC

GROWTH

RATE, AND

LIFE FOR LARVAE OF THE GASTROPOD

CREPIDULA

FORNICATA

(L.)

JAN A. PECHENIK Biology Department,

Tufts University, Medford, MA 02155, U.S.A.

Abstract: To determine if a relationship exists between rate of development and ability to delay metamorphosis, larvae of the prosobranch gastropod Crepidulu fomicatu (L.) were reared at 18 and 24 “C in the absence of metamorphosis-inducing cues. Average growth rates were determined directly by periodically subsampling and measuring larvae at both temperatures. Individual growth rates were estimated from size at, and date of, spontaneous metamorphosis. Larvae generally grew more quickly at the higher temperature, in terms of shell length, tissue dry weight, and protein content. Regardless of temperature, most (> 95%) larvae eventually metamorphosed spontaneously in clean glass dishes. Rapid growth was associated with an apparently shorter pre-competent period and earlier spontaneous metamorphosis, reflecting increased rates of differentiation. These results suggest a developmental mechanism through which selective pressures may operate to determine dispersal potential in this, and perhaps other species. Specifically, pre-competent and delay periods may be altered through selection for different rates of development. Demonstration of such a mechanism may ultimately increase our ability to predict the potential duration of planktonic existence of planktotrophic marine invertebrate larvae under a variety of environmental circumstances.

The length of

time that dispersive larvae of marine benthic invertebrates spend in the plankton prior to metamorphosis figures importantly in many ecological models of invertebrate reproductive patterns (e.g., Vance, 1973; Strathmann, 1974; Doyle, 1975; Pechenik, 1979; Caswell, 1981; Jackson & Strathmann, 1981). Larval life is divided into two parts: a pre-competent period, during which time larvae cannot be induced to metamorphose to adult form, and a competent period, during which time metamorphosis can be triggered by providing larvae with some chemical and/or physical component of the adult environment (Crisp, 1974; Scheltema, 1974; Hadfield, 1977). In the absence of such a cue, metamorphosis is delayed. The duration of this period of delayed metamorphosis (“the delay period”) seems generally to be finite, culminating either in spontaneous metamorphosis or in death (e.g., Wilson, 1948; Bayne, 1965; Hinegardner, 1969; Chia & Spaulding, 1972; Crisp, 1974; Pechenik, 1980; Kempf, 1981). The dispersal potential of a given larva is thus a function of both its rate of development to metamorphic competence and of the maximum length of time that metamorphosis can subsequently be delayed. Our ability to predict the maximum potential duration of planktonic development for any given species or individual is limited. Few studies have specifically set out to define 0022-0981/84/$03.00 0 1984 Elsevier Science Publishers B.V.

242

JAN A. PECHENlK

potential delay periods for marine ~nve~ebrate larvae, and the pbys~olo~~al and developmental bases for the ability to delay metamorphosis are little explored. In particular, the influence of environmental factors, such as temperature, on the potential duration of the deiay period has been examined only for larvae of the mussel, h4ytiltis eduir’s(Bayne, 1955). Rates of development prior to rnet~o~~~ competence are, in contrast, well known to be influenced by a variety ofenvironmental factors for a number of marine invertebrate species (reviewed by Pechenik, in press), and it seems likely that these factors will influence ability to delay metamorphosis in these species as well. Through what developmental mechanism might ability to delay metamorphosis be determined? I have suggested (Pechenik, 1980) that an eventual end to larval life may be “programmed” into development in at least some species. According to this hypothesis, slowdeveloping larvae should be able to delay metamorphosis for a longer period of time than larvae which develop more quickly. This hypothesis is based upon, and is supported by, interspe~i~~ comparisons of Iarvai delay periods and growth rates; there is an inverse correlation between the rate of larval development and the length of time that metamorphosis can be delayed (Pechenik, 1980; Jackson & Strathmarm, 1981). If delay periods are indeed influenced by rates of development, this inverse relationship should also be observable ~traspe~i~cally. The relationship between rate of development and potential length of larval life in a single species has not previously been examined experimentally. Although demonstration of such a relationship cannot prove the existence of a pre-programmed end to the maintenance of larval form, it can lend support to the hypothesis. In addition, documentation of such a relationship will help elucidate the degree to which environmental factors influence larval dispersal potential. In this paper, the hypothesis that duration of larval life is influenced by developmental rate is examined. Specifically, I witi consider the effect of temperature on both developmental rate and potential duration .of the planktonic period for larvae ofthe stipper shell snail, Crepi&l~ fimicata (L.), MATERIALS AND METHODS LARVAL REARING

CONDITIONS

Adult C. fomicata were collected from Woods Hole, Massachusetts, and maintained at 20 “C on a mixed diet of the naked flagellates Iso&rysis gdbana Parke Butch. and ~~~~~~~~~te&&32&2until larvae were released. This occurred after z 1 wk, on Day 0. All larvae used in the experiments described in this paper were released on a single day, but were not necessarily released by a single female and did not necessarily have the same father. Larvae were reared in sea water collected at Nahant, Massachusetts, and passed through a 0.4%pm filter. Larvae were fed a unialgal diet of ~~~~~~~~ga~~#~, since this alga supports rapid growth of Creptiula fimicata through metamorphosis (Pechenik, 1980).

DELAY OF METAMORPHOSIS

BY GASTROPOD LARVAE

243

A wide variation in larval growth rates was obtained by rearing larvae at 18, and 24 “C. Twenty-five larvae were reared in 45 ml of Isochtysis galbana suspension (0.55 larvae per ml) in each of five glass dishes at both temperatures. These individuals were used to monitor growth rates (non-destructive sampling) and the duration of the planktonic period in the absence of a metamorphic trigger. Other veligers were reared at the same larval density in larger glass dishes (200 ml I. galbana suspension). These individuals were used to monitor larval growth rates (destructive sampling) and to define the relationships between larval shell length, dry tissue weight, and protein content. An algal cell density of 18 x lo4 cells. ml- 1 was maintained in all cultures. This density is above the critical cell density for larvae of Crepidulafomicata~Pechenik, 1980), so that larvae were probably feeding at maximal rates in all experiments. Water and food were changed daily in all cultures, and all glassware was acid-rinsed before each use. MONITORING

LARVAL GROWTH AND METAMORPHOSIS

At l- to 2-day intervals, the shells of larvae in several of the large and/or small dishes at both temperatures were measured at 63 x using a dissecting microscope equipped with an ocular micrometer. Measurements made on larvae from the small dishes were made on living individuals, which were then returned to their respective containers. On two occasions all of the larvae in all of the containers, both large and small, were measured non-destructively. Larval growth rates did not differ significantly among containers at either 18 or 24 “C (P > 0.1, by one-way analysis of variance). Measurements made by destructive sampling of larvae taken from the large dishes, therefore, accurately reflected growth occurring in the smaller dishes as well. Metamorphosis of larval gastropods is signalled by the loss of the velum, the ciliated organ used for swimming and, in planktotrophic species, for suspension feeding as well (Scheltema, 1961; Fretter, 1972; Bonar & Hadfield, 1974). Metamorphosis, as thus defined, constitutes an irreversible, ecological transition from a planktonic to a benthic existence, as well as a morphological transition from larval to adult form. All larvae in the small dishes were inspected daily for resorption of the velum. Metamorphosed individuals were removed from containers, measured, and discarded. These measurements of newly metamorphosed individuals were later used to estimate individual growth rates, as explained below. A slight modification to the above protocol was made at 18 “C. On Day 12 of the experiment (i.e., 12 days after release of veligers from the adults), all of the individuals in the small dishes at 18 ‘C were measured and sorted into two groups based upon shell length. The mean shell length of larvae placed in the fast-growing group was 1032.4 pm (SD = 53.5 pm, N = 72), while the mean shell length of larvae placed in the slowgrowing group was 866.9 pm (SD = 69.2 pm, N = 48). Fast- and slow-growing individuals were distributed among five containers for each group. The volume of seawater/algal suspension in each container was adjusted periodically, so that the number of larvae per ml remained the same for all containers throughout the experiment.

JAN A. PECHENlK

244

Regression analyses were performed and analyzed following Kleinbaum & Kupper (1978). RELATIONSHIPSBETWEENLARVALSIZE, WEIGHT, AND

PROTEIN

CONTENT

Since shell length increments may not adequately reflect larval tissue growth, depending upon larval shell geometry and/or environmental conditions (Lucas & Costlow, 1979; Pechenik, 1980, and in press), measurements of biomass and protein content were made also. Periodically, some of the larvae from the large culture dishes were measured and homogenized in groups of 15-200 individuals in 150 ~1 of O.l”/, NP-40 for analysis of protein content, using a standard protein assay kit (Bio-Rad Laboratories). Absorbance at 595 nm was measured with a Bausch and Lomb Model 2000 spectrophotometer and protein content was calculated from a standard curve. Other larvae were removed periodically from the large 24 “C cultures and fixed in formalin for biomass determinations. These data were compared with similar data obtained previously for C.fimicala at 18-20 “C (Pechenik, 1980). The larval shells were removed by acidification (Pechenik, 1980), the tissue was rinsed with deionized water to remove salts, and the larvae were then transferred to pre-weighed containers. Due to the low weight of individual larvae, animals of similar shell length were grouped together to obtain a reliable weight estimation. Smaller larvae were weighed in larger groups than larger individuals. After 48 h of desiccation, weight determinations were made to the nearest 1 pg using a Cahn Model 21 electrobalance. RESULTS Shell lengths of larval C. fimicatu increased at different but constant rates at 18 and 24 “C (Fig. 1). Temperature altered the slope of the lines relating shell length to larval

0

2

4

5

8

‘Cl

I2

‘4

‘6,

DdY

Fig. 1. Rate of shell length increase as a function of rearing temperature: error bars represent I SD about the mean (IV = 40 for each point); larvae become competent to metamorphose at ~700 pm shell length.

DELAY OF METAMORPHOSIS

BY GASTROPOD

LARVAE

245

age (P < 0.05; t = 4.66, d.f. = 20), but not the intercepts (P > 0.10; t = 1.69, d.f. = 20). As had been found earlier (Pechenik, 1980), growth rates were maintained even after larvae became competent to metamorphose. The relationships between shell length and biomass (Fig. 2) and between shell length and protein content (Fig. 3) were not notice221 -$

F.

I8/

0

400

500

600

700

Shell

800

Length

900

1000

1100

(urn)

Fig. 2. Effect of temperature on the relationship between dry tissue weight and shell length: 0, 18-20 “C data (from Pechenik, 1978); 0,24 “C data; error bars represent 1 SD about mean shell length; each point was obtained using 6-38 individuals of similar shell length.

1

400

600

800 Mean

1000

1200

Shell

Length

1400

600

1800

(urn)

Fig. 3. Relationship between protein content and larval shell length at 18 “C (0) and at 24 “C (0): each point represents the demise of 15-200 individuals.

246

JAN A. PECHENIK

ably altered by temperature over the range 18-24 “C. Moreover, these relationships were linear at both temperatures (Figs. 2,3). Hence, growth rate at a given temperature (18 or 24 “C) was apparently constant in terms of both shell length, individual tissue biomass, and individual protein content. At 24 ‘C, larvae thus grew faster than they did at 18 ’ C in every respect: in terms of shell length increase, tissue biomass accumulation, and increase in protein content. Virtually all larvae eventually underwent spontaneous metamorphosis in the clean glass dishes at both temperatures. Of the 120 (18 “C) and 117 (24 “C) individuals accounted for at the end of the experiment, 95.5 and 98.4%, respectively, had metamorphosed. No larva died at a shell length of < 1600 pm, and the few mortalities that occurred were generally due to desiccation of (metamorphosed?) individuals crawling above the water level in the containers. Metamorphosis was first observed nearly 1 wk earlier at 24 “C than at 18 “C, and 50 % of the larvae metamorphosed by Day 24 at 24 ‘C; ,N2 additional wk were required for larvae reared at 18 “C to reach this level of spontaneous metamorphosis (Fig. 4). A level of 95 % spontaneous metamorphosis was attained about twice as quickly at the higher temperature (Fig. 4).

Fig. 4. Effect of temperature on the time required to reach spontaneous accounted for was 120 and 117 at 18 and 24°C

metamorphosis: respectively.

number

of larvae

The fast-growing larvae at 18 “C metamorphosed sooner than did slower-growing larvae at the same temperature (P < 0.05; Z = 4.59, d.f. = 41) (Fig. 5). These data were analyzed following an arcsine transformation of percentage metamorphosis (Sokal & Rohlf, 1969). The slow-growing individuals required approximately 1 wk longer to reach 50% spontaneous metamorphosis than did the faster-growing group (Fig. 5). Variability in the duration of planktonic life was greater for larvae reared at 18 ‘C than for those reared at 24 “C; i.e., the number of days over which spontaneous metamor-

DELAY OF METAMORPHOSIS

BY GASTROPOD

247

LARVAE

phosis occurred was greater at the lower temperature (Fig. 5). This variability is correlated with an apparent influence of temperature on the variability of larval growth rate, growth rate being more variable at 18 “C than at 24 “C (Fig. 6).

20--

y=200x-2502 r=O996

IO -.

OkH:~:~::i:'::::::::~:::!:::;;;;:~:::i 0 25 30 35 40 45

50

55

Fig. 5. Effect of differences in growth rate on the time required to reach spontaneous metamorphosis at 18 “C: the fast-growing group (0) contained 72 individuals; the slow-growing group (0) contained 48 individuals.

No spontaneous metamorphosis occurred before the geometry of shell growth had transformed from the spiral pattern characteristic of pre-competent larvae to the linear form characteristic of adults (Pechenik, 1980). Shells of larvae which delay metamorphosis resemble brimmed hats, and are easily recognizable in laboratory culture or in field plankton samples (see Fig. 3E of Thiriot-Quievreux & Scheltema, 1982). Larvae reared at 24 “C tended to form shell brims at smaller shell lengths in comparison with larvae reared at 18 “C (Table I). That is, some individuals at 24 “C developed brims at a smaller size than any individuals at 18 “C. In addition, some individuals reached a larger size without forming shell brims at 18 ‘C than did any individuals at 24 ‘C. The development of individual larvae over time could not be monitored, since larvae were reared in batch cultures. Thus, mean shell length at brim formation cannot be calculated for these experiments. Spontaneous metamorphosis occurred over a wide range of shell lengths at both temperatures (Figs. 7, 8). At 18 “C, for example, shell lengths of metamorphosed individuals ranged between w 900 and 2300 pm. The smallest individuals collected

0 LI::!! 0 350

i! 550

450

Mean

~) 650 Larval

I 750 Shell

I)

::

Ii

850 Length

950

II 1050

1150

(urn)

Fig. 6. Effect of temperature on variability of larval growth rates: N = 25 individuals for each point. TABLE I Influence of temperature on the larval development of Crepidula fomicaru: growth rates were determined from the data given in Fig. 1.

Size at Average Temp. (“C)

formation

(pm.day-‘)

spontaneous

metamorphosis

Size at brim

growth rate

(pm)

(pm)

N

(mean f SD)

N

18

45.3

843-931

162

1601 f 285

120

24

71.7

802-883

182

1473 f 184

117

2300

1

2100 -

. .

19-c Y = 17.34x + 970.49

r =0.58 N= 102

. .

.

.*

l

. .

s

v

1900-

1700 -

l500-

.

1300-

IIOO-

9001

r’,~I~~~i~I~I~l~I~I~~~I

1

0 I8 22 26 30 34 38 42 46 50 54 58 62 DQY

Fig. 7. Size at spontaneous metamorphosis

as a function of time at 18 “C.

DELAY OF METAMORPHOSIS

BY GASTROPOD LARVAE

249

within 24h of spont~~us

met~o~hosis had shell lengths of 956 pm at 18 “C and 1172 pm at 24 “C. The largest individuals collected within 24 h of metamorphosis had shell lengths of 2372 pm (18 “C) and 1989 pm (24 “C).

0 14

20

25

30

35

40

Day

Fig. 8. Size at spontaneous metamorphosis as a function of time at 24 “C.

The mean shell length of larvae measured within 24 h of spontaneous metamorphosis differed significantly as a function of rearing temperature (Table I, P -c0.01, oneway analysis of variance). At both temperatures, the relationship between the day on which au individual metamorphosed and its shell length within 24 h of metamorphosis was statistically significant (Figs. 7, 8; the critical value of Yat the 1% probability level is z%0.25). One goal of this study was to examine the relationship between growth rate and dispersal potential. Ideally, this would be done by following the growth rate and time of metamorphosis of individually-reared larvae. This is currently being done. The large number of larvae used in the present experiments, however, made such an approach unfeasible. Instead, the form of the relations~p between rate of ~di~du~ growth (pm shell length added per day, which also reflects growth in terms of biomass and protein content - Figs. 2, 3), and the duration of planktonic life was examined using the data summarized in Figs. 7 and 8. The date of metamorphosis and the shell size of newly metamorphosed individuals are known. Larval life began on Day 0 for all individuals and larval shell length on Day 0 is highly uniform among individuals (mean i: 1 SD = 393.2 f 19.5 pm, N = 27). Individual, average rates of shell growth are thus easily estimated. I have chosen to consider total duration of planktonic development in this calculation, since this gives the best indication of dispersal potential. The rela~onship between growth rate and duration of the delay period will have the same form, and can be estimated by assuming that larvae become competent to metamorphose at 700 pm shell length as discussed later.

250

JAN A. PECHENIK

Total length of planktonic life varied between 18 and 6 1 days at 18 ’ C, and between 14 and 39 days at 24 “C (Figs. 7, 8). For each temperature, a statistically significant (P < 0.01) inverse correlation exists between time to spontaneous metamorphosis and estimated larval growth rate (Fig. 9). The correlation coefficients (r) at 18 and 24 “C

”

0

24’c

0

m

r=-072

ki ; 0

I5

--__c__+__

!

25

35

Growth Fig. 9. Time until spontaneous

_-_t_+___-

45

Rate

metamorphosis as a function calculation of growth

55

65

75

( urn/day) of larval growth rate.

rate: see text for details

of

are - 0.42 and - 0.46, respectively. However, < 20% of the variation in when the larvae metamorphosed is explained by variation in growth rate (r’ < 0.20 at both temperatures). Note that the times of spontaneous metamorphosis for slower-growing larvae at 24 “C fall well within the range of values obtained for larvae growing at comparable rates at the lower temperature. Similarly, the fastest-growing individuals at 18 “C metamorphosed within the range of times recorded for larvae growing at comparable rates at 24 oC. The effect of temperature seems primarily to have been to extend the range of growth rates observed. The correlation coefftcient relating growth rate to time until spontaneous metamorphosis for all of the data combined is r = - 0.72 (N = 237) (Fig. 9). Even so, only x 50% of the variation in duration of planktonic development in the laboratory was explained by variation in growth rate (r’ = 0.52). The relationship between size at metamorphosis and individual growth rate is significant at both 18 and 24 “C (r = 0.394 and 0.332, respectively; P < 0.05by F-test) (Fig. 10). The regression coefficients (slopes) are significantly different from zero for both temperatures (P< 0.05). That is, at both 18 and 24 “C, there was a tendency for

DELAY OF METAMORPHOSIS

BY GASTROPOD LARVAE

251

faster-growing individuals to metamorphose at a larger size. The slopes of the two lines (Fig. 10) are not significantly different from each other at the 5% level, but acceptance of the null hypothesis is borderline (2 = 1.88, critical value = 1.97). However, the two 25007

2300--z .s .I z k .i 0 f r

ZIOO--

‘8-C r=0.42

1

.

. .

1900--

1700.-

I

25

Growth

35

Rote

45

55

65

(urn/day)

Fig. 10. Size at metamorphosis as a function of individual growth rate at 18 “C (0) and 24°C (0).

lines differ in elevation, i.e. magnitude of the relationship between growth rate and size at metamorphosis is altered by temperature (analysis of covariance, Nie etal., 1975; P < 0.01, F = 50.3, d.f. = 1,219).

DISCUSSION

Ability to delay metamorphosis seems to vary widely among species, both in the length of time that metamorphosis can be delayed, and in the eventual fate of larvae in the continued absence of a triggering cue. To some extent, these differences may reflect whether or not the larvae can feed in the plankton. The delay potential of lecithotrophic larvae must be limited by their inability to feed on particulates (e.g., Lucas etal., 1979; Pechenik et al., 1979; Satterlie & Case, 1979). However, wide variability is apparent even among species with planktotrophic larvae, even within a single group such as the Gastropoda. For example, larvae of Aplysia juliana can delay metamorphosis (23-30 “C) for up to 288 days (Kempf, 1981). In contrast, the larvae of Bittium altematum and Zlyanassa obsoleta delayed metamorphosis (18-20 ‘C) for only x 2 months (Pechenik, 1980), the larvae of Crepidulafomicata (18-24 “C) delayed

252

JAN A. PECHENIK

for z 3 to 4 wk (Pechenik, 1980, and present study), and the larvae of Doridella obscura had a competent period of 5 2 wk at 25 “C (Perron & Turner, 1977). Three observations require discussion: (1) metamorphosis of larval Crepidula fomicata was not delayed indefinitely at either 18 or 24 “C; (2) virtually all individuals eventually underwent successful, spontaneous metamorphosis - larval mortality was exceptionally low; (3) faster-growing larvae generally underwent spontaneous metamorphosis earlier than did slower-growing individuals, both between and within temperatures. There is no obvious reason why the competent period of a feeding gastropod larva should not extend indefinitely (Crisp, 1974). Veliger larvae of C.fornicata remain active and show no sign of energy imbalance during the delay period, i.e., the time between the onset of competence and the occurrence of spontaneous metamorphosis (Pechenik, 1980). These larvae continue to grow in terms of shell length, tissue biomass, and individual protein content after becoming competent to metamorphose (Pechenik, 1980, and present study). Moreover, I have collected large, brimmed larvae from the plankton in both surface and deep-water tows near Woods Hole, Massachusetts (Pechenik, 1978), confirming the laboratory finding that the larvae remain active swimmers during the extended competent period. Yet, the larvae of C. fimicata and of many other species eventually seem to reach an endpoint at which time the larvae either die or metamorphose spontaneously in the absence of a triggering cue. Species-specific differences in delay capability should result in differences in dispersal capability and are likely to be adaptive: they may be related, at least in part, to differences in the way that adult habitats are distributed (Wilson, 1948; Bayne, 1965; Crisp, 1974; Pechenik, 1980), and/or to the frequency with which larvae are dispersed away from favorable habitats (Jackson & Strathmann, 1981). The limited number of interspecific comparisons which can presently be made suggest that the amount of time required to reach the “larval endpoint” may be dependent upon rate of development (Pechenik, 1980; Jackson & Strathmann, 1981). Data presented here for C. fomicata are consistent with the hypothesis of a predetermined end to the maintenance of larval form and function. Faster growth was associated with a shorter planktonic life. This was true between temperatures (Figs. 1,4) and within temperatures (Figs. 5, 9). The maximum duration of the dispersal period is thus related to the rate at which C. fornicata progresses through its developmental program; an individual which develops rapidly will quickly become competent to metamorphose and will reach the larval endpoint disproportionately sooner than an individual which develops more slowly (Fig. 11). It may eventually be possible to predict the potential duration of planktonic life for planktotrophic larvae under different environmental circumstances by simply knowing the environmental effect on growth. Variations in larval growth rate explained at most only about 50% of the variability seen in duration of larval life of C. fornicata, despite the fact that the inverse correlation between the two parameters is statistically significant. Development has two components : growth and differentiation. Growth is a measure of increased body size and/or weight and/or cell number, whereas differentiation is morphological or physiological

DELAY OF METAMORPHOSIS

253

BY GASTROPOD LARVAE

change that reflects temporal shifts in gene expression (Highnam, 1981). Growth and differentiation are generally correlated. For example, certain developmental events are clearly correlated with rates of growth in larvae of the insect, Munduca sex& (Nijhout & Williams, 1974). As shown for amphibians by Smith-Gill & Berven (1979), however,

Adult

/

larval Stage

endpoint

of

Development

f

/ ’

onset of competence *

x Death

/

/

Larva

Time

-

Fig. 11. Model of the developmental mechanism potentially explaining the relationship between developmental rate and capacity for delay of metamorphosis: maximum duration of planktonic life is limited by the rate of development towards a pre-programmed end to the maintenance of larval form and function.

the fastest-growing individuals in a population are not necessarily differentiating the most rapidly, even though rates of growth and differentiation are positively correlated. The low value for r2 obtained in my experiments may, in large part, reflect a weak correlation between rates of growth and differentiation. To some extent, the low rz value may also reflect error associated with the estimation of growth rates from sizes of newly-metamorphosed individuals. For Crepzihla fornicata, both size at onset of shell brim formation and size at spontaneous metamorphosis were significantly altered as a function of temperature-dependent growth rate (Table I). This suggests that growth and differentiation rates are influenced independently by changes in temperature. This hypothesis is further supported by the finding that the relationship between size at spontaneous metamorphosis and magnitude of individual growth rate is altered by temperature change. Further experiments monitoring the growth and differentiation of individual larvae are currently in progress. Calculation of actual delay periods at the two temperatues is problematic. The major difficulty is the determination of the point at which an individual larva becomes competent to metamorphose. In the absence of a morphological trait which is func-

254

JAN A. PECHENIK

tionally associated with the onset of competence, the only reliable way to demonstrate that larvae are competent is to successfully induce metamorphosis. It then becomes impossible to monitor delay periods for these individuals. Brim formation may provide a useful indicator of metamorphic competence for larvae of C. fimicatu. This is currently being examined. Estimates of delay periods at each temperature can be obtained by assuming that C. fornicatu larvae reach competence at 700 pm shell length (Werner, 1955). I have routinely induced metamorphosis of C. fornicata larvae in the size range 700-750 pm at 18-25 “C (Pechenik, 1978, 1980, and additional unpublished experiments). Mean size at emergence from egg capsules was 393.2 pm. Average rates of larval growth during the first lo-12 days of development at 18 and 24 “C were 45.3 and 71.7pm.dayy’, respectively (Fig. 1). Competence was therefore reached after ~4.3 days at 24 “C and 6.8 at 18 “C (Fig. 1). Variation in actual size at competence should introduce little error into calculation of the pre-competent period since larval growth rates are high. Fifty percent of the larvae at 24 “C underwent spontaneous metamorphosis by Day 24, and the comparable result was obtained by Day 40 for larvae at 18 “C (Fig. 4). The median duration of the delay period was thus z 19.5 days at 24 “C and 33 days at 18 “C. These values are in close agreement with the range of delay periods (17-30 days) previously estimated for larvae of C.fbrnicatu reared at 18-20 “C (Pechenik, 1980). The average daily growth rate represented in Fig. 9 is 29.6 (SD = 7.3) pm I day ’ at 18 “C and 42.5 (SD = 7.2) pm. day-’ at 24 “C. These rates were estimated from measurements of shell lengths at or shortly after the time ofmetamorphosis, as explained earlier. The ranges of individual growth rates estimated in this manner were 16.8-57.2pm.dayy’ at 18 “C and 30.1-65.5 ym.day-- ’ at 24 “C. However, mean daily growth rates actually measured by non-destructive subsampling during the first lo-12 days after larval release were higher at both 18 and 24 “C; 45.3 pm. day- ’ at 18 oC and 7 1.7 pm . day - ’ at 24 ‘C (Fig. 1). The implication is that larval growth rates declined later in development. Algal cell densities may not have been sufficiently high to permit maximal growth of large larvae, which clear phytoplankton from suspension at a more rapid rate than do smaller individuals (Pechenik, 1980). Growth of individual larvae will be followed throughout development to resolve this issue. Unlike the larvae of the few prosobranch gastropods which have been studied so far (Pechenik, 1980), the larvae of opisthobranch gastropods do not increase in shell length during the period of delayed metamorphosis, although differentiation continues after shell growth ceases (Switzer-Dunlap & Hadfield, 1977; Bickell & Chia, 1979). This confirms an independence between the processes of growth and differentiation. Thus, cessation of growth by opisthobranch larvae during delay of metamorphosis, followed by either death or spontaneous metamorphosis, is not incompatible with the hypothesis of a developmentally programmed larval endpoint. Cessation of growth need not guarantee a capacity for prolonged delay of metamorphosis, provided that differentiation continues. On the other hand, a species which curtails further differentiation once

DELAY OF METAMORPHOSIS

BY GASTROPOD LARVAE

255

becoming competent to metamorphose, whether or not it continues to grow, may be capable of delaying metamorphosis indefinitely, assuming that nutritional and metabolic needs continue to be met. The nature of the larval endpoint is unknown. Conceivably, there may be some relationship to the phenomena of senescence and programmed cell death (Holliday & Pugh, 1975; Pechenik, 1980; Highnam, 1981). Alternatively, the larval endpoint could reflect shifts in the sensitivity of sensory receptors, the responsiveness of target cells to a circulating hormonal factor, and/or shifts in the concentration of such a factor in the body fluids. As yet, no hormonal involvement in the metamorphosis of molluscan larvae has been demonstrated (Hadtield, 1978; Highnam, 1981). A precise understanding of the nature of the larval endpoint must await an increased understanding of the genetic and physiological components of the processes through which larvae of marine invertebrates become competent and are triggered to initiate metamorphosis by contact with appropriate stimuli (Hadfield, 1978; Highnam, 1981). The approaches that have been used so successfully to examine the developmental basis of insect and amphibian metamorphosis (e.g., Kollros, 1961; Riddiford, 1975; Safranek & Williams, 1980) are not applicable to marine invertebrate larvae due to the very small size of these larvae. The indirect approach used here may provide a basis for further insights into an otherwise elusive problem. The apparent existence of a pre-programmed larval endpoint in at least some marine invertebrate species offers a mechanism through which natural selection could operate to bring about differences in delay capability within, and, perhaps, between species. This could be achieved simply by selecting for differences in rates of individual development. ACKNOWLEDGEMENTS

I wish to acknowledge J. LeBlanc and L. Hutton for excellent technical assistance and statistical advice, respectively. This work was supported by an NIH Biomedical Research Grant administered through Tufts University, and by NSF Grant OCE8121643. The manuscript benefitted from the constructive criticisms of Dr. R. Doyle and an anonymous reviewer.

REFERENCES BAYNE,B. L., 1965. Growth and delay of metamorphosis of the larvae ofMyrilus edu1i.s(L.). Opheliu, Vol. 2, pp. l-47. BICKELL, L.R. & F.-S. CHIA, 1979. Organogenesis and histogenesis in the planktotrophic veliger of Doridella steinbergne (0pisthobranchia:Nudibranchia). Mar. Biol., Vol. 52, pp. 291-313. BONAR,D. B. & M. G. HADFIELD, 1974. Metamorphosis of the marine gastropod Phestiflu sibogae Bergh (Nudibranchia:Aeolidacea). I. Light and electron microscopic analysis oflarval and metamorphic stages. J. Exp Mar. Biol. Ecol., Vol. 16, pp. 227-255. CASWELL,H., 1981. The evolution of “mixed” life histories in marine invertebrates and elsewhere. Am. Nut., Vol. 117, pp. 529-536.

256

JAN A. PECHENIK

CHIA, F.-S. & J.G. SPAULDING, 1972. Development and juvenile growth of the sea anemone, Teaha crassicornis. Biol. Bull. (Woods Hole, Mass.), Vol. 142, pp. 206-218. CRISP, D. J., 1974. Factors influencing the settlement of marine invertebrate larvae. In, Chemoreception in marine organisms, edited by P. T. Grant & A.M. Mackie, Academic Press, New York, pp. 177-265. DOYLE,R., 1975. Settlement of planktonic larvae: a theory of habitat selection in varying environments. Am. Nar., Vol. 109, pp. 113-126. FRETTER,V., 1972. Metamorphic changes in the velar musculature, head and shell of some prosobranch veligers. J. Mar. Biol. Assoc. U.K., Vol. 52, pp. 161-177. HADFIELD,M. G., 1977. Chemical interactions in larval settling of a marine gastropod. In, Marine natural products chemistry, edited by D. J. Faulkner & W. H. Fenical, Plenum, New York, pp. 403-413. HADFIELD,M.G., 1978. Metamorphosis in marine molluscan larvae: an analysis of stimulus and response. In, Settlement and metamorphosis of marine invertebrate larvae, edited by F.-S. Chia & M.E. Rice, Elsevier/North-Holland, Amsterdam, pp. 165-175. HINEGARDNER,R.T., 1969. Growth and development of the laboratory cultured sea urchin. Biol. Bull. (Woods Hole, Mass.), Vol. 137, pp. 465-475. HIGHNAM,K.C., 1981. A survey of invertebrate metamorphosis. In, Metamorphosis: a problem in developmental biology, 2nd edition, edited by L. I. Gilbert & E. Frieden, Plenum, New York, pp. 43-73. HOLLIDAY,R. & J.E. PUGH, 1975. DNA modification mechanisms and gene activity during development. Science, Vol. 187, pp. 226-232. JACKSON,G.A. & R.R. STRATHMANN,1981. Larval mortality from offshore mixing as a link between precompetent and competent periods of development. Am. Nat., Vol. 118, pp. 16-26. KEMPF, S. C., 1981. Long-lived larvae of the gastropod Aplysiujuliana: do they disperse and metamorphose or just slowly fade away? Mar. Ecol. Progr. Ser., Vol. 6, pp. 61-65. KLEINBAUM,D.G. & L.L. KUPPER, 1978. Applied regression analysis and other multivariable methods, Wadsworth Publ. Co., 556 pp. KOLLROS,J. J., 1961. Mechanisms ofamphibian metamorphosis: hormones. Am. Zool., Vol. 1, pp. 107-l 14. LUCAS,J. S. & J. S. COSTLOW,JR., 1979. Effects of various temperature cycles on the larval development of the gastropod mollusc Crepidulu fornicata. Mar. Biol., Vol. 5 1, pp. 11l-l 17. LUCAS,M. I., G. WALKER,D. L. HOLLAND& D. J. CRISP, 1979. An energy budget for the free-swimming and metamorphosing larvae of Balanus bulanoides (Crustacea: Cirripedia). Mar. Biol., Vol. 55, pp. 221-229. NIE, N.H.. C. H. HULL, I. G. JENKINS,K. STEINBRENNER 62 D. H. BENT, 1975. Stlrtistirui puckage,for the social sciences, second ed., McGraw-Hill, New York, 675 pp. NIJHOUT, H.F. & C.M. WILLIAMS, 1974. Control of moulting and metamorphoses in the tobacco hornworm, Munduca sextu (L.): cessation ofjuvenile hormone secretion as a trigger for pupation. .I. Exp. Biol., Vol. 61, pp. 493-501. PECHENIK,J.A., 1978. Growth and energy balance during the development and delay of metamorphosis of larval gastropods. Ph.D. thesis, University of Rhode Island, U.S.A., 153 pp. PECHENIK,J. A.,1979. Role of encapsulation in invertebrate life histories. Am. Nat., Vol. 114, pp. 859-870. PECHENIK,J. A., 1980. Growth and energy balance during the larval lives of three prosobranch gastropods. J. Exp. Mar. Biol. Ecol., Vol. 44, pp. l-28. PECHENIK,J.A., in press. Environmental influences on larval growth and survival. In, Reproduction of marine invertebrates, Vol. 9, edited by A.C. Giese & J. S. Pearse, Academic Press, New York. PECHENIK,J. A., F. E. PERRON& R. D. TURNER,1979. The role of phytoplankton in the diets of adult and larval shipworms, Lyrodus pedicellutus (Bivalvia : Teredinidae). Estuuries, Vol. 2, pp. 58-60. PERRON,F. E. & R. D. TURNER,1977. Development, metamorphosis, and natural history of the nudibranch Doridella obscura Verrill (Corambidae : Opisthobranchia). J. Exp. Mar. Biof. Ecol., Vol. 27, pp. 171-185. RIDDIFORD,L. M., 1975. Juvenile hormone-induced delay of metamorphosis of the viscera of the cecropia silkworm. Biol. Bull. (Woods Hole, Mass.), Vol. 148, pp. 429-439. SAFRANEK,L. & C.M. WILLIAMS, 1980. Studies of the prothoracicotropic hormone in the tobacco hornworm, Munduca sexta. Biol. Bull. (Woods Hole, Mass.), Vol. 158, pp. 141-153. SAXTERLIE,R. A. & J. F. CASE, 1979. Development of bioluminescence and other effector responses in the pennatulid coelenterate Renillu kiillikeri. Biol. Bull. (Woods Hole, Mass.), Vol. 157, pp. 506-523. SCHELTEMA,R. S., 1961. Metamorphosis of the veliger larvae of Nassarius obsoletus (Gastropoda) in response to bottom sediment. Biol. Bull. (Woods Hole. Mass.), Vol. 120, pp. 92-109.

DELAY OF METAMORPHOSIS

BY GASTROPOD

LARVAE

257

SCHELTEMA, R. S., 1974. Biological interactions determining larval settlement of marine invertebrates. Thalassia Jugosl., Vol. 10, pp. 263-296. SOKAL, R. R. & F.J. ROHLF, 1969. Biometry. Theprincigles andpractice ofstatistics in biologicalresearch, W.H.

Freeman Co., San Francisco, 776 pp. BERVEN, 1979. Predicting amphibian metamorphosis. Am. Mat., Vol. 113, pp. 563-585. STRATHMANN, R., 1974. The spread of sibling larvae of sedentary marine invertebrates. Am. Nat., Vol. 108, pp. 29-44. SWITZER-DUNLAP,M. & M. G. HADFIELD, 1977. Observations on development, larval growth and metamorphosis of four species of Aplysiidae (Gastropoda : Opisthobranchia) in laboratory culture. J. Exp. Mar. Biol. Ecol., Vol. 29, pp. 245-261. THIRIOT-QUI~VREUX,C. & R. S. SCHELTEMA,1982. Planktonic larvae of New England gastropods. V. Bittium alternatum, Ttiphora nigrocincta, Cerithiopsis emersoni, Lunatia heros and Crepidula plana. Malacologia, Vol. 23, pp. 37-46. VANCE, R.R., 1973. On reproductive strategies in marine benthic invertebrates. Am. Nat., Vol. 107, pp. 339-352. WERNER,B., 1955. Uber die Anatomie, die Entwicklung und Biologie des Veligers und der Veliconcha von Crepidula fornicata. Helgol. Wiss. Meeresunters., Vol. 5, pp. 169-217. WILSON, D.P., 1948. The relationship of the substratum to the metamorphosis of Ophelia larvae. J. Mar. Biol. Assoc. U.K., Vol. 27, pp. 723-759. SMITH-GILL, S.J. & K.A.