Annual Reproductive Cycle of the Green

S. K. Meidel R. E. Scheibling

Annual reproductive cycle of the green sea urchin, Strongylocentrotusdroebachiensis, in differing habitats in Nova Scotia, Canada

Received: 20 May 1997 /Accepted: 21 January 1998

Abstract We monitored the reproductive cycle ofStrongylocentrotus droebachiensis (OF Muller) betweenApril 1993 and August 1995 in kelp beds, barrengrounds and grazing fronts at both a wave-exposed anda sheltered site along the Atlantic coast of Nova Scotia.Gonad index and histological analyses showed thatS. droebachiensis has an annual reproductive cycle that issynchronous across sites and habitats, and between fe-males and males. Spawning occurs in March/April ofeach year but a small proportion of sea urchins in thestudy populations also spawned in fall 1995. Duringmost of the year, sea urchins in kelp beds and grazingfronts have a higher gonad index than those in barrengrounds. Gonad indices also tended to be higher at thewave-exposed than the sheltered site. Interannual vari-ability in peak gonad index was significant in the barrengrounds at the wave-exposed site and in the grazingfront at the sheltered site. The gametogenic cycle ischaracterized by six stages based on the abundance ofnutritive and germinal/gametic cells. Nutritive phago-cytes are abundant after spawning and replaced by in-creasing numbers of germinal and gametic cells as thegametogenic cycle progresses. The temporal patterns ofabundance of each cell type were similar among habitatsindicating that the gonads were qualitatively similardespite large dierences in gonadal mass. The quantityof gut contents (ratio of food volume to body volume)was similar among habitats, but the quality (percentageof organic material) tended to be higher in kelp beds andgrazing fronts than in barren grounds suggesting thatdierences in gonad index of S. droebachiensis in dif-ferent habitats are related to dierences in diet. The highdensity of sea urchins in grazing fronts combined with

their high fecundity suggests that they make the greatestcontribution, per unit area, to the overall larval pool.

Introduction

The green sea urchin, Strongylocentrotus droebachiensis,is the dominant herbivore in the shallow, rocky, subtidalzone in eastern Canada (Miller and Mann 1973; Mann1977). Along the Atlantic coast of Nova Scotia, large-scale fluctuations in population size of S. droebachiensiscause dramatic changes in the state of the shallow sub-tidal ecosystem (Mann 1977; Wharton and Mann 1981;Miller 1985; Scheibling 1986). When sea urchins are inlow abundance, kelp beds (mainly Laminaria longicrurisand L. digitata and various understorey algae) flourishin the rocky subtidal zone. Sea urchins in kelp beds areusually cryptic and sparsely distributed. They functionmainly as detritivores consuming drift algae in crevicesand under boulders (Mann 1985). As sea urchins in-crease in number, they begin to aggregate along the edgeof kelp beds forming fronts which destructively grazethe kelp (Breen and Mann 1976; Lang and Mann 1976;Wharton 1980). These grazing fronts can advance atrates of 1 to 4 m per month creating extensive barrengrounds denuded of fleshy macroalgae (Breen and Mann1976; Scheibling et al. 1994).

The mechanisms leading to sea urchin populationincreases are poorly understood, but may include spo-radic recruitment events (Hart and Scheibling 1988;Scheibling 1996) and/or migration (Foreman 1977;Scheibling et al. in preparation). The formation of densegrazing fronts may initiate positive feedback mecha-nisms that drive a population outbreak. For example,increased fecundity due to consumption of kelp (Vadas1977; Larson et al. 1980), or increased fertilization ratedue to the proximity of spawning individuals (Penning-ton 1985), may result in increased larval production. Ifthe advance of the fronts is uninterrupted (e.g. by massmortality because of disease or harvesting), the subtidalecosystem will shift from the kelp bed to the barren

Marine Biology (1998) 131: 461478 Springer-Verlag 1998

Communicated R.J. Thompson, St. Johns

S.K. Meidel (&) R.E. ScheiblingDepartment of Biology, Dalhousie University,Halifax, Nova Scotia, B3H 4J1, Canada

ground state over several years (Breen and Mann 1976;Mann 1977). Sea urchins may persist long after thedisappearance of kelp beds but rates of growth andreproduction decrease as they adjust to lower foodavailability (Lang and Mann 1976; Wharton and Mann1981; Johnson and Mann 1982; Chapman and Johnson1990).

Strongylocentrotus droebachiensis has an annual re-productive cycle with a major spawning period (as evi-denced by a decline in gonad index) in late winter orearly spring (Cocanour and Allen 1967; Himmelman1978; Falk-Petersen and Lnning 1983; Keats et al.1984; Munk 1992). Some spawning also has been ob-served in summer and fall o Newfoundland (Keats et al.1987). Numerous studies have shown that food quan-tity and quality strongly influence reproduction ofS. droebachiensis and other sea urchins (e.g. Lasker andGiese 1954; Ebert 1968; Lawrence 1975; Vadas 1977;Larson et al. 1980). The greater gonad index ofS. droebachiensis in kelp beds than barren grounds(Lang and Mann 1976; Wharton 1980; Johnson andMann 1982; Keats et al. 1984; Sivertsen and Hopkins1995) is generally attributed to dierences in foodavailability between these two habitats (Lang and Mann1976; Sivertsen and Hopkins 1995). However, few in-vestigators have included gut content analysis in theirstudies, and usually only the occurrence of particularfood items is recorded (Himmelman and Steele 1971;Chapman 1981; Himmelman and Nedelec 1990). Con-sequently, there is little quantitative information tocompare the amounts and type of food consumed by seaurchins in kelp beds versus barren grounds.

Wave exposure is another factor that may directly orindirectly influence the reproduction of sea urchins at asite. For example, the supply of drift algae may begreater at wave-exposed sites due to increased wave ac-tion which dislodges and transports plants (Rogers-Bennett et al. 1995). However, Ebert (1968) and Gonor(1973a) found that Strongylocentrotus purpuratus at ex-posed sites had reduced gonad indices compared tothose at sheltered sites. Ebert (1968) attributed this dif-ference to a higher cost of repair for broken spines at theexposed site, leaving less energy available for repro-duction.

In the present study, we compare the reproduction ofsubpopulations of Strongylocentrotus droebachiensis inkelp beds and barren grounds, and in grazing fronts atthe ecotone between these two habitats, at both a wave-exposed and a sheltered site in Nova Scotia. We use bothgonad index and histological methods to quantify thereproductive cycle and to examine the eects of habitatand site on maturation and spawning. Also, we comparegut contents of sea urchins in the dierent habitats andsites to relate dierences in reproductive patterns toquantity and quality of consumed food. Finally, wecombine data on reproduction with other populationcharacteristics to examine the relative contribution ofsea urchins in kelp beds, grazing fronts, and barrengrounds to the overall larval pool.

Materials and methods

Study sites and sea urchin subpopulations

We studied the reproductive cycle of Strongylocentrotus droebac-hiensis (OF Muller) at two sites along the southwestern shore ofNova Scotia: Little Duck Island (4422N; 6411W), a wave-ex-posed island at the mouth of Mahone Bay, and Mill Cove(4435N; 643W), a sheltered cove in St. Margarets Bay. At LittleDuck Island, the substratum consisted of basaltic bedrock inter-sected by ridges and grooves. At Mill Cove, the underlying graniticbedrock was covered with rocks and boulders. At both sites, thestudy areas were at a depth of 6 to 9 m.

We compared sea urchins from kelp beds and adjacent barrengrounds, and from grazing fronts at the interface between the twohabitats. Kelp beds at both sites consisted of a dense canopy ofLaminaria longicruris with an understorey of branching (e.g. Cer-amium rubrum, Plumaria plumosa) and foliose algae (e.g. Chondruscrispus, Palmaria palmata), and articulated coralline algae (Coral-lina ocinalis). At Little Duck Island, kelp plants were relativelyshort with narrow and rued blades, a morphology associated withhigh wave exposure (Gerard and Mann 1979). At Mill Cove, kelpdensity was lower and the plants were longer, wider, and thinner.Barren grounds at both sites were dominated by encrusting coral-line algae (mainly Phymatolithon laevigatum, Lithothamnion glaci-ale) with scattered patches of ephemeral filamentous algae (mainlyDesmarestia viridis) appearing in summer/fall. Barren grounds alsoreceived input of drift algae (mainly kelp) from the adjacent kelpbeds. The grazing front at the interface of the kelp bed and barrengrounds was characterized by kelp stipes (stripped of blades) andarticulated corallines, which were the last erect macroalgae to beconsumed by the sea urchins.

At both sites, sea urchin density and mean size diered in spaceand time. In the kelp beds at both sites, sea urchins were sparselydistributed throughout the study period (mean density:

tion. Total body wet weight and gonad wet weight were measuredwith an electronic balance (0.01 g accuracy). Gonad index wascalculated [(gonad wet weight/total body wet weight) 100] to givea percentage. Sex was determined by examining a gonad smearunder a compound microscope. Horizontal test diameter wasmeasured with vernier calipers (0.05 mm accuracy).

Temporal patterns in gonad index of female and male sea ur-chins were compared across habitats (kelp bed, grazing front,barren grounds) using three-way analysis of variance (ANOVA)with Date (March 1994 to August 1995, when sea urchins weresampled concurrently in all three habitats), Habitat, and Sex asfixed factors. Gonad indices for each sex at the peak of the re-productive cycle were compared between years using one-wayANOVA (grazing front and barren grounds at Little Duck Island,1993 to 1995) or t-tests (kelp bed at Little Duck Island, grazingfront and barren grounds at Mill Cove, 1994 and 1995; a missedsampling interval for the kelp bed at Mill Cove at the peak of thereproductive cycle in 1994 precluded statistical analysis in thishabitat). Gonad index at the peak of the reproductive cycle(March/April 1995) and after spawning was completed (June 1995)was compared between sites and sexes, and among habitats (allclassified as fixed factors) by three-way ANOVA. We classified Siteas a fixed factor because the two study sites were chosen to rep-resent dierent degrees of exposure to wave action. Raw data werearcsine transformed to remove heterogeneity of variance as indi-cated by Cochrans C-test ( p < 0.05). Because sample sizes variedbetween sites, dates, habitats and sexes, we used Type III sums ofsquares, and carried out post-hoc comparisons using the GT2-method (Sokal and Rohlf 1995).

To examine changes in gonad index with body size in Strong-ylocentrotus droebachiensis and to confirm that the gonad indexof adult sea urchins within the size range used in our study wasindependent of test diameter, we sampled 66 to 75 sea urchinsbetween 14.3 and 74.9 mm in each habitat at the peak of the re-productive season in 1995 (late March/early April). In S. droebac-hiensis, the development of gonad index with increasing testdiameter can be described with a logistic growth model (Munk1992). We related gonad index to size using the following function:

Y Y0MY0 M Y0ekMd15 ; 1

where Y is gonad index, Y0 is gonad index in immature sea urchins(given a small positive value, 0.1), M is the asymptotic gonad index,k is a constant, and d is test diameter. In all cases the logistic modelprovided a better fit to our data than a straight-line regression. Weused linear regression techniques to analyse the relationship be-tween gonad index (arcsine transformed) and adult body size (35 to50 mm) in S. droebachiensis at the peak of the reproductive cycle.In 50% of samples collected at the peak of the reproductive cycle,a few individuals (usually

Results

Spatial and temporal patterns in gonad index

Strongylocentrotus droebachiensis displays a distinct an-nual cycle of reproduction as indicated by temporalchanges in gonad index between 1993 and 1995 (Fig. 1).Most spawning occurred in March/April of each year,resulting in a sharp drop in gonad index. In the kelp bedand the grazing front at Mill Cove in 1995, the peak

gonad index declined more slowly and spawning mayhave extended into May. The overall cycle is relativelysynchronous across sites and habitats, and also betweenfemales and males. At each site, there was a significant

Fig. 1 Strongylocentrotus droebachiensis. Mean gonad index (per-centage of total body wet weight, SD) for female, male or unsexedsea urchins (35 to 50 mm test diameter) at Little Duck Island andMillCove between April 1993 and August 1995 in the kelp bed, the grazingfront and the barren grounds. Means are based on 2 to 17 sea urchins

464

interaction between the eects of sampling date andhabitat on gonad index (Table 1). Post-hoc comparisons(GT2-test) showed that the gonad index in the barrenswas significantly lower than in the kelp bed and/or thegrazing front on all dates except September 1994 atLittle Duck Island, and on all dates except in March andOctober 1994 at Mill Cove. There also was a significantinteraction between the eects of sampling date and sexon gonad index (Table 1). At Little Duck Island, femaleshad a significantly higher gonad index than males at thepeak of the gonad index cycle in April 1994 and March1995, and males had a significantly higher index thanfemales in September 1994. At Mill Cove, females alsohad a higher gonad index than males in April 1994, andmales had a higher index than females in December1994.

At Little Duck Island, the peak gonad index in thebarren grounds increased significantly from 1993 to 1995for each sex (females: F2;14 5:87, p 0:014; males:F2;11 11:39, p 0:002), but there were no significantinterannual dierences in peak gonad index in either thegrazing front (females: F2;19 0:90, p 0:422; males:F2;21 1:33, p 0:285) or the kelp bed (females:t9 1:57, p 0:150; males: t22 1:90, p 0:071). AtMill Cove, peak gonad index in the barren grounds didnot dier significantly between 1994 and 1995 (females:t11 0:41, p 0:692; males: t6 0:98, p 0:365), butthe gonad index in the grazing front was significantlyhigher in 1995 than in 1994 (females: t25 2:28,

p 0:031; males: t19 2:90, p 0:009). The peak gonadindex immediately prior to spawning in 1995 did notdier significantly between sites (F1;105 1:60,p 0:209) but diered consistently between habitats atboth sites (i.e. mean gonad index was highest in the kelpbed, lowest in the barren grounds; F2;105 33:34,p < 0:001). Gonad index also was consistently higher forfemales than males (F1;105 10:91, p 0:001): there wasno significant interaction between site, habitat and sex.The post-spawning gonad index (June 1995) was sig-nificantly higher at Little Duck Island than at Mill Cove(F1;84 18:06, p < 0:001). It was consistently higher inthe kelp bed and grazing front than in the barrengrounds at both sites (F2;84 41:33, p < 0:001), and didnot dier significantly between females and males(F1;84 1:74, p 0:190): there was no significant inter-action between site, habitat and sex.

The relationship between the gonad index and testdiameter of Strongylocentrotus droebachiensis just beforespawning (Fig. 2) indicates that the development ofmacroscopic gonads begins at a size of 15 mm in allhabitats at both sites. Gonad index increases rapidlybetween 25 and 35 mm and then tends towards an as-ymptote that is determined by habitat. Linear regressionconfirmed that there was no relationship between gonadindex and test diameter over the size range that weused to monitor the reproductive cycle (35 to 50 mm)(Table 2). There were no signs of reproductive senes-cence in large individuals up to 75 mm.

Table 1 Strongylocentrotus droebachiensis. Three-way ANOVA ofthe eects of Date, Habitat, and Sex on gonad index (arcsinetransformed) and GT2 post-hoc comparisons of the simple eectsof Habitat (sexes pooled) and Sex (habitats pooled) at each date at

Little Duck Island and Mill Cove [Date: Mar/Apr 1994 to Aug1995; Habitat: kelp bed (KB), grazing front (GF ), barren grounds(BG ); Sex: female ( f ), male (m); NS not significant; *p < 0:05;**p < 0:01; ***p < 0:001; nd no data]

Test Little Duck Island Mill Cove

ANOVA df MS F p df MS F p

Date 12 609.46 76.66 BG NSJun 1994 GF>BG NS KB>GF>BG NSJul/Aug 1994 KB>BG NS KB, GF>BG NSSep 1994 NS fBG NSOct 1994 KB>GF, BG NS NS NSNov 1994 KB>GF, BG NS KB, GF>BG NSDec 1994 KB>GF, BG NS KB, GF>BG fGF, BG NSFeb/Mar 1995 KB, GF>BG f>m KB, GF>BG NSMar/Apr 1995 KB, GF>BG NS KB>GF>BG NSMay 1995 KB, GF>BG NS KB, GF>BG NSJun 1995 KB, GF>BG NS KB, GF>BG NSAug 1995 KB, GF>BG NS KB>GF>BG NS

465

Gametogenic cycle

The gametogenic cycles of female and male Strong-ylocentrotus droebachiensis were characterized by sixmaturity stages as illustrated by representative micro-graphs (Fig. 3). In Stage I (recovering) gonadal acini arefilled with storage cells (nutritive phagocytes), and small

numbers of germinal cells (oocytes in females, sperm-atocytes in males) are present along the acinal walls(Fig. 3a, g). In Stage II (growing), nutritive phagocytesdecrease in abundance and are replaced by increasingnumbers of oocytes or spermatocytes (Fig. 3b, h). InStage III (premature), nutritive phagocytes further de-crease in abundance and the first mature gametes (ova

Fig. 2 Strongylocentrotusdroebachiensis. Relationship be-tween gonad index and test di-ameter (14.3 to 74.9 mm) inMarch (Mill Cove) and April(Little Duck Island) 1995 in thekelp bed, the grazing front, andthe barren grounds. The plottedline represents the fit of Eq. 1 toeach set. Parameter values for M(asymptotic gonad index), and k(a constant) are given for eachrelationship. Y0 (gonad index injuveniles) equals 0.1 in all cases (nsample size; r2 coecient of de-termination)

466

or spermatozoa) begin to accumulate in the lumen(Fig. 3c, i). In Stage IV (mature), most of the lumen isoccupied by mature gametes, and nutritive phagocytesare reduced to a thin layer along the acinal wall (Fig. 3d,j). In Stage V (partly spawned), the lumen is emptied asmature gametes are shed but not yet replaced to anygreat extent by nutritive phagocytes (Fig. 3e, k). InStage VI (spent), some relict oocytes/ova or spermato-zoa may be present in the lumen, which is accumulatinga growing layer of nutritive phagocytes (Fig. 3f, l).

The gametogenic cycle of Strongylocentrotusdroebachiensis was approximately synchronous betweensites and across habitats for both males and females,although individuals could be found in two or threedierent maturity stages on most dates (Figs. 4, 5). Afterspawning in spring, females remained in the recoveringstage (Stage I) for 2 to 4 months before moving into thegrowing stage (Stage II) during the summer (Fig. 4). Bylate summer or early fall, most females had entered thepremature stage (Stage III) where they remained untillate winter or early spring when they became fully mature(Stage IV). Females proceeded rapidly through the partlyspawned (Stage V) and spent stages (Stage VI) andstarted a new gametogenic cycle a few weeks afterspawning. At Mill Cove, one partly spawned female wasfound in September in the kelp bed (Fig. 3e). Males of S.droebachiensis showed a similar pattern of maturation asfemales, although the periodicity was less pronounced(Fig. 5). After spawning, most males entered the recov-ering and growing stages in early or mid-summer,although up to 30% of males in some habitats (LittleDuck Island, barren grounds; Mill Cove, kelp bed)remained in the spent stage until late summer. At LittleDuck Island, most males entered the premature stage inlate fall while at Mill Cove 25% of males were still inthe growing stage in February. Most males were fullymature in late winter or early spring, and proceededthrough the partly spawned and spent stages within 1 to 2months of spawning before starting a new game-togenic cycle. At Mill Cove, one mature male wasfound in October in both the grazing front and the barrengrounds (Fig. 3j), and one partly spawned male wasfound in November in the grazing front (Fig. 3k).

Changes in gonadal microstructure during maturation

The proportion (by cross-sectional area of a gonadalacinus) of nutritive phagocytes in ovaries of females ofStrongylocentrotus droebachiensis showed a distinct an-nual cycle that was synchronous across sites and habi-tats (Fig. 6). After the major spawning period in March/April, the proportion of nutritive phagocytes increasedrapidly within 2 months. As gametogenesis proceeded,the proportion of nutritive phagocytes progressivelydecreased to a minimum just prior to the next majorspawning period. The proportion of nutritive phago-cytes in the ovaries diered significantly between dates atboth sites (Table 3), and it was significantly lower in thekelp bed than in the grazing front at Little Duck Island(T -test). Mean oocyte area increased throughout thematuration cycle and reached a maximum just prior tospawning, when it decreased sharply as large oocytesmatured into ova and newly produced oocytes weresmall (Fig. 6). At Little Duck Island, mean oocyte areadiered significantly between dates but not betweenhabitats (Table 3). At Mill Cove, there was a significantinteraction between date and habitat: mean oocyte areawas significantly lower in the grazing front than in thebarren grounds in February and the kelp bed in March.While oocytes were present at all times, ova first ap-peared in late winter and were lost at spawning (Fig. 6).There were no significant dierences in mean ova areabetween months or habitats at either site (Table 3). Therelative abundance of ova, and the proportions ofoocytes, nutritive phagocytes, and unoccupied lumenwere used to quantify the maturity stages of females(Table 4).

Males of Strongylocentrotus droebachiensis showedthe same temporal pattern in the proportion of nutritivephagocytes in the gonads as females (Fig. 7). The pro-portion of nutritive phagocytes increased rapidly afterspawning and then progressively decreased until the nextmajor spawning period. The proportions of sperm-atocytes and spermatozoa (Fig. 7) showed a reciprocalpattern of abundance relative to nutritive phagocytes.After spawning, the proportion of spermatocytes in-creased to a maximum in early winter and remained atthat level until the next spawning. The proportion ofspermatozoa dropped sharply after spawning and re-mained low during the summer, increasing in fall andwinter to a maximum at the peak of the reproductivecycle. At Little Duck Island, there was a significant in-teraction between the eects of date and habitat on theproportions of all three cell types in the testes (Table 5).The proportion of nutritive phagocytes was significantlyhigher and the proportion of spermatozoa significantlylower in the barren grounds than in the kelp bed and/orgrazing front in October 1994 and May 1995. The pro-portion of spermatocytes also was significantly lower inthe barren grounds than in the kelp bed and grazingfront in May 1995. At Mill Cove, there was a significanteect of date on the proportions of both nutritivephagocytes and spermatocytes but no significant eect

Table 2 Strongylocentrotus droebachiensis. Results of linear re-gression analysis of gonad index (arcsine transformed) on testdiameter (34.5 to 52.2 mm) (n sample size; r2 coecient of de-termination; p probability)

Site,habitat

Size range (mm) n r2 p

Little Duck IslandKelp bed 35.249.8 17 0.152 0.122Grazing front 34.550.9 13 0.035 0.541Barren grounds 35.050.2 15 0.104 0.240

Mill CoveKelp bed 34.752.2 28 0.033 0.650Grazing front 35.050.3 24 0.008 0.685Barren grounds 34.750.0 20 0.057 0.310

467

468

of habitat (Table 5). Also at Mill Cove, there was asignificant interaction between the eects of date andhabitat on the proportion of spermatozoa which wassignificantly higher in the barren grounds than in thegrazing front in February 1995. The proportions ofspermatocytes, spermatozoa, nutritive phagocytes, andunoccupied lumen were used to quantify the maturitystages of males (Table 6).

Sex ratio

Sex ratios of Strongylocentrotus droebachiensis did notdeviate significantly from 1:1 (v2-test, p > 0:05) in anyhabitat at either site with the exception of the kelp bed atLittle Duck Island, where males were more abundantthan females (148 males, 115 females; v2 4:141,p < 0:05). Samples in which >10% of urchins could notbe sexed were excluded from analysis. Three hermaph-rodites were observed at Mill Cove (one from eachhabitat), which represented 0.35% of sea urchins sam-pled at that site n 862 and 0.15% of the total sam-pled at both sites n 1968.

Gut content analysis

The food quantity index of Strongylocentrotus droebac-hiensis was temporally variable in all habitats at bothsites but tended to be lowest in late summer and earlyfall (Fig. 8). At Little Duck Island the index increased inthe kelp bed and grazing front after spawning (March/April) in 1995. At both sites, there was a significant in-teraction between the eects of date and habitat on thefood quantity index (Table 7). At Little Duck Island, theindex was significantly lower in the barren grounds thanin the kelp bed and/or the grazing front in fall 1994 andspring 1995 (5 out of 11 dates; GT2-test) and signifi-

cantly lower in the kelp bed than in the grazing frontand/or barren grounds in late summer and fall 1994 andJune 1995 (6 out of 11 dates). At Mill Cove, the foodquantity index was significantly lower in the barrengrounds than in the kelp bed and/or grazing front inJune and December 1994, and in late winter/early spring1995 (5 out of 12 dates).

The food quality index of Strongylocentrotusdroebachiensis (Fig. 8) was consistently high in the kelpbed and grazing front and more variable but generallylower in the barren grounds at both sites. As with thefood quantity index, there also was a significant inter-action between the eects of date and habitat on thefood quality index (Table 7). At Little Duck Island, thefood quality index was significantly lower in the barrengrounds than in the kelp bed and/or grazing front inspring and fall 1994, and spring and summer 1995 (8 outof 13 dates). At Mill Cove, this was the case in summerand winter 1994 and throughout 1995 (9 out of 13dates).

Discussion

Reproductive cycle

Strongylocentrotus droebachiensis on the Atlantic coastof Nova Scotia exhibits a distinct annual reproductivecycle with a major spawning period in early spring. Thecycle was relatively synchronous between habitats dif-fering in food quality and quantity, and between sitesdiering in wave exposure. Previous studies have showna similar cycle of gonad index for S. droebachiensis inMaine (Cocanour and Allen 1967), Newfoundland(Himmelman 1978; Keats et al. 1984), and Norway(Falk-Petersen and Lnning 1983). Histological analysisalso indicated a similar progression of non-gametic andgametic cells as previously described for females ofS. droebachiensis (Falk-Petersen and Lnning 1983) andfor both sexes of other strongylocentrotids (e.g. Fuji1960; Chatlynne 1969; Gonor 1973a, b). Nutritivephagocytes were most abundant at the beginning of thereproductive cycle and were subsequently replaced byincreasing numbers of germinal and gametic cells (oo-cytes and ova in females, spermatocytes and spermato-zoa in males).

The general synchrony of reproduction in all habitatssuggests that the annual reproductive cycle is controlledby factors other than food, possibly temperature and/orphotoperiod (e.g. Gonor 1973a). Individual sea urchins,however, usually occurred in two or three gametogenicstages at any one time, with the greatest variabilitypresent during the spawning period. Such variation,which also has been documented in other sea urchins(Crapp and Willis 1975; Bernard 1977; Byrne 1990; Kinget al. 1994), is likely related to individual dierences inthe acquisition and allocation of energy reserves to ga-metogenesis. To our knowledge, our study is the first toquantitatively document changes in cell type abundance

b

Fig. 3 Strongylocentrotus droebachiensis. Histology of ovaries (af )and testes (gl). (a) Stage I: recovering ovary with nutritive phagocytes(NP) filling lumen; few small oocytes (Oc) along acinalwall. (b) Stage II:growing ovary with more abundant and larger oocytes along acinalwall. (c) Stage III: premature ovary with many oocytes accumulatingin lumen; nutritive phagocyte layer reduced. (d) Stage IV: matureovary filled with ova (O); nutritive phagocytes are reduced to thinlayer along acinal wall (Nu nucleus). (e) Stage V: partly spawnedovary with spaces vacated by spawned ova. (f ) Stage VI: spent ovarywith relict ova and few new oocytes; nutritive phagocyte layerincreasing in thickness. (g) Stage I: recovering testes with nutritivephagocytes (NP) filling lumen; thin layer of spermatocytes (Sc) alongacinal wall. (h) Stage II: growing testes with spermatocyte layerincreasing in thickness. (i) Stage III: premature testes with sperma-tozoa (Sz) accumulating in lumen; nutritive phagocyte layer reduced.(j) Stage IV: mature testes filled with spermatozoa; nutritivephagocytes are reduced to thin layer along acinal wall. (k) Stage V:partly spawned testes with spaces vacated by spawned spermatozoa (Llumen). (l) Stage VI: spent testes with nutritive phagocytes almostfilling lumen; scattered spermatocytes along acinal wall (Scale bars:100 lm)

469

in the gonads of Strongylocentrotus droebachiensis, andthus serves as a benchmark for future histologicalstudies of the reproductive cycle of this species.

A more gradual decline in gonad index during thespring spawning period at Mill Cove compared to LittleDuck Island suggests that spawning was more pro-tracted or occurred somewhat later at the former site. In

the northwestern Atlantic, spawning of Strong-ylocentrotus droebachiensis is triggered by phytoplank-ton blooms (Himmelman 1975; Starr et al. 1990, 1992,1993) which vary in space and time. Dierences intemperature or hydrodynamic regimes between our sitesmay have influenced the occurrence of phytoplanktonblooms and hence the timing of sea urchin spawning.

Fig. 4 Strongylocentrotusdroebachiensis. Frequencies (%)of females in Stages I to VI of thereproductive cycle at Little DuckIsland and Mill Cove betweenJune 1994 and June 1995 in thekelp bed, the grazing front, andthe barren grounds. Stage I:recovering, Stage II: growing,Stage III: premature, Stage IV:mature, Stage V: partly spawned,and Stage VI: spent. Numbersabove bars indicate sample size

470

Histological analysis revealed that a small proportionof the population of Strongylocentrotus droebachiensis atMill Cove spawned in fall. Although the incidence ofsummer and fall spawning is low, it corroborates ob-servations by Keats et al. (1987) of spawning ofS. droebachiensis in June, July and September in barrengrounds in Newfoundland. Because of the low numberof sea urchins that may spawn in the summer or fall, it is

unlikely that these events would contribute much to theoverall pool of larvae produced each year.

Spatial and interannual variation in gonad index

Gonad indices of Strongylocentrotus droebachiensisgenerally were higher in the kelp bed and grazing front

Fig. 5 Strongylocentrotusdroebachiensis. Frequencies (%)of males in Stages I to VI of thereproductive cycle at LittleDuck Island and Mill Covebetween June 1994 and June1995 in the kelp bed, the grazingfront, and the barren grounds.For stage description see Fig. 4.Numbers above bars indicatesample size

471

than in the barren grounds. This pattern is consistentwith previous studies contrasting the gonad index of thisspecies (Lang and Mann 1976; Keats et al. 1984;Scheibling and Stephenson 1984; Sivertsen and Hopkins1995) or other strongylocentrotids (Gonor 1973a; Pearse1980) between kelp beds and barren grounds, and pre-sumably is related to dierences in food availability (seeBetween habitat variation in food consumption, be-low). Several studies have shown that laminarian kelpsare a preferred food of S. droebachiensis which supportshigh rates of growth and reproduction (Vadas 1977;

Keats et al. 1984; Lemire and Himmelman 1996; Minorand Scheibling 1997). Our histological analysis indicatedthat the gonads were qualitatively similar between hab-itats (in terms of the proportions of dierent cell types)despite large dierences in gonadal mass. In contrast,Minor and Scheibling (1997) found that females ofS. droebachiensis fed kelp (Laminaria longicruris) ad li-bitum in the laboratory had significantly more nutritivephagocytes in their gonads than those fed kelp only oneday per week, and suggested that the higher rationprovided additional reserves for gametogenesis. How-

Fig. 6 Strongylocentrotusdroebachiensis. Mean (SD)relative area (percentage ofcross-sectional area of gonadalacini) of nutritive phagocytes,and mean (SD) absolute ar-eas of oocytes and ova offemale sea urchins in the kelpbed, grazing front and barrengrounds at Little Duck Islandand Mill Cove between June1994 and May 1995. Means arebased on 2 to 12 sea urchins

472

ever, greater between-diet dierences in gonad produc-tion in the laboratory study may account for this dis-parity.

The peak gonad index increased between 1993 and1995 in the barren grounds at Little Duck Island, whichmay reflect a reduction in intraspecific competition forfood after the mass mortality in October 1993 (Scheib-ling and Hennigar 1997). There were no interannualdierences in peak gonad index in the kelp bed orgrazing front during this period, suggesting that foodsupply (mainly kelp) was not limiting reproduction ineither of these two habitats. Other studies comparinggonad indices over several years also have shown inter-annual dierences in peak gonad index (Himmelman1978; Keats et al. 1984; Munk 1992) which in some caseswere related to dierences in food supply (Keats et al.1984).

Gonad indices usually were higher at the wave-ex-posed site, Little Duck Island, than at the sheltered site,

Mill Cove. In contrast, Ebert (1968) and Gonor (1973a)found that Strongylocentrotus purpuratus had lower go-nad indices at exposed sites than at sheltered sites, whichEbert attributed to greater energy allocation to spinerepair at exposed sites. In both studies, however, dif-ferences in wave exposure were confounded with dif-ferences in food abundance, which was lower (Ebert1968) or higher (Gonor 1973a) at the sheltered site.

Between habitat variation in food consumption

At both sites, the quantity of the gut contents ofStrongylocentrotus droebachiensis was lowest in latesummer and early fall when gonad indices also were low.This suggests a decrease in feeding rate at this timewhich is consistent with observations of sea urchin be-haviour at Little Duck Island and Mill Cove during theperiod of study: sea urchins in the grazing front becameless aggregated and grazed less actively on kelp in thelate summer and fall (Scheibling unpublished data). Gutcontents at Little Duck Island also were relatively low atthe peak of the reproductive cycle but increased afterspawning. Previous studies of S. droebachiensis (Vadas1977; Himmelman 1980; Keats et al. 1983; Himmelmanand Nedelec 1990) and congeneric species (Lawrenceet al. 1965; Ebert 1968; Vadas 1977) also have shown adecline in feeding rate in late summer/early fall with aminimum around the peak of the reproductive cycle.The large dierences in the abundance of macroalgalfood resources between kelp beds and barren groundswere not reflected in large dierences in the quantity ofgut contents of sea urchins from these habitats. How-ever, as sea urchins decrease gut evacuation rate whenfood is scarce (Lasker and Giese 1954; Propp 1977), thequantity of gut contents in barren grounds may notadequately reflect the level of food consumption or

Table 3 Strongylocentrotus droebachiensis. Two-way ANOVA ofthe eects of Date and Habitat on proportions of nutritive pha-gocytes, and absolute areas of oocytes and ova of females at LittleDuck Island and Mill Cove (Date: Jun, Oct, Dec 1994, and Feb/

Mar, Mar/Apr and May 1995; Habitat: kelp bed, grazing front,barren grounds; NS not significant; *p < 0:05; **p < 0:01;***p < 0:001)

Source Little Duck Island Mill Cove

df MS F p df MS F p

Nutritive phagocytesDate 5 3895.49 78.63

availability. Therefore, a significant dierence may existin the quantity of food consumed between barrengrounds and kelp beds which we were unable to detect.

At both sites, food quality in terms of organic ma-terial tended to be lower in the barren grounds than inthe kelp bed or at the grazing front. Vadas (1977) foundthat food quality is more important than quantity forreproduction in Strongylocentrotus droebachiensis,which may explain the lower gonad index of urchins inbarren grounds. Nevertheless, sea urchins in barrengrounds are able to obtain sucient nutrients for growthand reproduction owing to their generalist diet and

ability to locate and consume drift algae such as kelps(Himmelman and Steele 1971; Lawrence 1975; Vadas1977; Mann et al. 1984; Keats et al. 1984). Meidel andScheibling (1998) found that the growth rate of adult seaurchins did not dier significantly among habitats atLittle Duck Island, although it was somewhat slower inthe barren grounds than in the kelp bed or grazing frontat Mill Cove. If sea urchins channel a similar proportionof energy into growth in all habitats, reduced energyintake in barren grounds should result in reduced re-production. Also, foraging costs may be higher in barrengrounds where individuals tend to move greater dis-

Fig. 7 Strongylocentrotusdroebachiensis. Mean (SD)relative abundance (percentageof cross-sectional area of go-nadal acini) of nutritivephagocytes, spermatocytes, andspermatozoa of male sea ur-chins in the kelp bed, grazingfront and barren grounds atLittle Duck Island and MillCove between June 1994 andMay 1995. Means are based on3 to 8 sea urchins

474

tances than in kelp beds or grazing fronts (Mattison et al.1977; Harrold and Reed 1985; Scheibling unpublisheddata), which would further reduce the amount of energyavailable for reproduction.

Sex ratio and sexual dierences in gonad index

The sex ratio of Strongylocentrotus droebachiensis ap-proximated 1:1 in all cases except for the kelp bed atLittle Duck Island, where males accounted for a slightlyhigher proportion of the population (56%). Munk(1992) also reported a slight bias towards males in onepopulation of S. droebachiensis in Alaska (59%), but aslight bias towards females in another population(56%). Biased sex ratios have also been reported forcongeneric species (Gonor 1973c; Bernard 1977), al-though gonochoric echinoderms such as strong-ylocentrotids typically have a sex ratio of 1:1 (Lawrence1987). The incidence of hermaphroditism in our studywas very low and similar to that found in other go-nochoric sea urchins (Bernard 1977; Lawrence 1987;Byrne 1990; King et al. 1994).

At the peak of the reproductive cycle in spring 1995,females had a higher gonad index than males at bothsites, which is consistent with previous studies ofStrongylocentrotus droebachiensis (Munk 1992; Minor

and Scheibling 1997) but not other strongylocentrotids(Bennett and Giese 1955; Bernard 1977). After spawn-ing, gonad indices of both sexes dropped to the sameminimal levels, indicating that females released a largerproportion (10.5%) of their body weight as gametesthan males (8.1%).

Spatial variation in zygote production

A number of studies have shown that fertilization rate insea urchins and other echinoderms is positively relatedto fecundity (which generally increases with increasingbody size) and population density (e.g. Pennington 1985;Levitan et al. 1992; Levitan 1995 and references therein).Adults of Strongylocentrotus droebachiensis in barrengrounds have low fecundity (because of their small sizeand low gonad index), while those in kelp beds have ahigh gonad index but are sparsely distributed. In con-trast, sea urchins in grazing fronts are both highly ag-gregated and much larger than those in barren groundsand kelp beds (Scheibling et al. 1994; Scheibling un-published data) and therefore are expected to have thehighest fertilization rate and produce the greatest num-ber of zygotes per unit area of bottom. During ourstudy, sea urchins at Little Duck Island had higher fe-cundity and occurred at higher densities than those at

Table 5 Strongylocentrotus droebachiensis. Two-way ANOVA ofthe eects of Date and Habitat on proportions of nutritive pha-gocytes, spermatocytes and spermatozoa of males at Little Duck

Island and Mill Cove (Date: Jun, Oct, Dec 1994, and Feb/Mar,Mar/Apr and May 1995; Habitat: kelp bed, grazing front, barrengrounds; NS not significant; *p < 0:05; **p < 0:01; ***p < 0:001)

Source Little Duck Island Mill Cove

df MS F p df MS F p

Nutritive phagocytesDate 5 4465.30 57.86

Table 7 Strongylocentrotus droebachiensis. Two-way ANOVA ofthe eects of Date and Habitat on food quantity and quality index,as well as GT2 post-hoc comparisons of the simple eects of Ha-bitat at each date at Little Duck Island and Mill Cove [Date: food

quantity, Jun 1994 to Aug 1995; food quality, Apr/May 1994 toAug 1995; Habitat: kelp bed (KB), grazing front (GF), barrengrounds (BG); NS not significant; *p < 0:05; **p < 0:01;***p < 0:001; nd no data]

Test Little Duck Island Mill Cove

ANOVA df MS F p df MS F p

Food quantity indexDate 10 601.93 29.88 KB, BG NS NS NSDec 1994 GF>BG>KB NS KB, GF>BG KB>GF>BGJan 1995 nd nd NS KB>GF>BGFeb/Mar 1995 NS KB, GF>BG GF>BG KB, GF>BGMar/Apr 1995 KB, GF>BG KB>GF>BG KB, GF>BG KB, GF>BGMay 1995 GF>BG KB>BG KB, GF>BG KB, GF>BGJun 1995 GF>KB>BG KB, GF>BG NS KB>GF>BGAug 1995 NS KB, GF>BG BG>KB KB, GF>BG

Mill Cove. Consequently, sea urchins at Little DuckIsland probably also experienced higher fertilizationsuccess and produced more zygotes per unit area.

Acknowledgements We thank T. Balch, I. Dempsey, M. Gedamke,and A. Hennigar for help in the field, W. Blanchard for advice onstatistical analyses, and especially A. Cameron for help with his-tological techniques. We also thank J. Himmelman, P. Yund andan anonymous reviewer for helpful comments on an earlier draft.SKM was supported by a Dalhousie University Graduate Schol-arship and an Izaak Walton Killam doctoral scholarship. Thestudy was funded by a research grant from the Natural Science andEngineering Research Council of Canada (NSERC), and a sciencesubvention grant from NSERC and the Department of Fisheriesand Oceans, Canada, to RES.

References

Bennett J, Giese AC (1955) The annual reproductive and nutri-tional cycles in two western sea urchins. Biol Bull mar biol Lab,Woods Hole 109: 206237

Bernard FR (1977) Fishery and reproductive cycle of the red seaurchin, Strongylocentrotus franciscanus, in British Columbia.J Fish Res Bd Can 34: 604610

Breen PA, Mann KH (1976) Destructive grazing of kelp by seaurchins in eastern Canada. J Fish Res Bd Can 33: 12781283

Byrne M (1990) Annual reproductive cycles of the commercial seaurchin Paracentrotus lividus from an exposed intertidal and asheltered subtidal habitat on the west coast of Ireland. Mar Biol104: 275289

Chapman ARO (1981) Stability of sea urchin dominated barrengrounds following destructive grazing of kelp in St. MargaretsBay, eastern Canada. Mar Biol 62: 307311

Chapman ARO, Johnson CR (1990) Disturbance and organizationof macroalgal assemblages in the Northwest Atlantic. Hydro-biologia 192: 77121

Chatlynne LG (1969) A histochemical study of oogenesis in the seaurchin, Strongylocentrotus purpuratus. Biol Bull mar biol Lab,Woods Hole 136: 167184

Cocanour B, Allen K (1967) The breeding cycles of a sand dollarand a sea urchin. Comp Biochem Physiol 20: 327331

Crapp GB, Willis ME (1975) Age determination in the sea urchinParacentrotus lividus (Lamarck), with notes on the reproductivecycle. J exp mar Biol Ecol 20: 157178

Ebert TA (1968) Growth rates of the sea urchin Strongylocentrotuspurpuratus related to food availability and spine abrasion.Ecology 49: 10751091

Falk-Petersen I-B, Lnning S (1983) Reproductive cycles of twoclosely related sea urchin species, Strongylocentrotus droebac-hiensis (OF Muller) and S. pallidus (GO Sars). Sarsia 68: 157164

Foreman RE (1977) Benthic community modification and recoveryfollowing intensive grazing by Strongylocentrotus droebachien-sis. Helgolander wiss Meeresunters 30: 468484

Fuji A (1960) Studies on the biology of the sea urchin. I. Superficialand histological gonadal changes in gametogenic process of twosea urchins, Strongylocentrotus nudus and S. intermedius. BullFac Fish Hokkaido Univ 11: 114

Gerard VA, Mann KH (1979) Growth and production of Lami-naria longicruris (Phaeophyta) populations exposed to dierentintensities of water movement. J Phycol 15: 3341

Gonor JJ (1973a) Reproductive cycles in Oregon populations of theechinoid, Strongylocentrotus purpuratus (Stimpson). I. Annualgonad growth and ovarian gametogenic cycles. J exp mar BiolEcol 12: 4564

Gonor JJ (1973b) Reproductive cycles in Oregon populations ofthe echinoid, Strongylocentrotus purpuratus (Stimpson). II.Seasonal changes in oocyte growth and in abundance ofgametogenic stages in the ovary. J exp mar Biol Ecol 12: 6578

Gonor JJ (1973c) Sex ratio and hermaphroditism in Oregon in-tertidal populations of the echinoid Strongylocentrotus purpur-atus. Mar Biol 19: 278280

Harrold C, Reed DC (1985) Food availability, sea urchin grazing,and kelp forest community structure. Ecology 66: 11601169

Hart MW, Scheibling RE (1988) Heat waves, baby booms, andthe destruction of kelp beds by sea urchins. Mar Biol 99:167176

Himmelman JH (1975) Phytoplankton as a stimulus for spawn-ing in three marine invertebrates. J exp mar Biol Ecol 20:199214

Himmelman JH (1978) Reproductive cycle of the green sea urchin,Strongylocentrotus droebachiensis. Can J Zool 56: 18281836

Himmelman JH (1980) The role of the green sea urchin, Strong-ylocentrotus droebachiensis, in the rocky subtidal region ofNewfoundland. In: Pringle JD, Sharp GJ, Caddy JF (eds)Proceedings of the workshop on the relationship between seaurchin grazing and commercial plant/animal harvesting. Cantech Rep Fish aquat Sciences 954: 92119

Himmelman JH, Nedelec H (1990) Urchin foraging and algalsurvival strategies in intensely grazed communities in easternCanada. Can J Fish aquat Sciences 47: 10111026

Himmelman JH, Steele DH (1971) Foods and predators of thegreen sea urchin Strongylocentrotus droebachiensis in New-foundland waters. Mar Biol 9: 315322

Johnson CR, Mann KH (1982) Adaptations of Strongylocentrotusdroebachiensis for survival on barren grounds in Nova Scotia.In: Lawrence JM (ed) Echinoderms: Proceedings of the Inter-national Conference, Tampa Bay. AA Balkema, Rotterdam,pp 277283

Keats DW, Hooper RG, Steele DH, South GR (1987) Field ob-servations of summer and autumn spawning by Strong-ylocentrotus droebachiensis, green sea urchins, in easternNewfoundland. Can Fld Nat 101: 463465

Keats DM, Steele DH, South GR (1983) Food relations and shortterm aquaculture potential of the green sea urchins (Strong-ylocentrotus droebachiensis) in Newfoundland. Tech Rep memlUniv Newf mar Sci Res Lab 24: 124

Keats DW, Steele DH, South GR (1984) Depth-dependent repro-ductive output of the green sea urchin, Strongylocentrotusdroebachiensis (O.F. Muller), in relation to the nature andavailability of food. J exp mar Biol Ecol 80: 7791

King CK, Hoegh-Guldberg O, Byrne M (1994) Reproduction cycleof Centrostephanus rodgersii (Echinoidea), with recommenda-tions for the establishment of a sea urchin fishery in New SouthWales. Mar Biol 120: 95106

Lang C, Mann KH (1976) Changes in sea urchin populations afterthe destruction of kelp beds. Mar Biol 36: 321326

Larson BR, Vadas RL, Keser M (1980) Feeding and nutritionalecology of the sea urchin Strongylocentrotus drobachiensis inMaine, USA. Mar Biol 59: 4962

Lasker R, Giese AC (1954) Nutrition of the sea urchin, Strong-ylocentrotus purpuratus. Biol Bull mar biol Lab, Woods Hole106: 328340

Lawrence JM (1975) On the relationships between marine plantsand sea urchins. Oceanogr mar Biol A Rev 13: 213286

Lawrence JM (1987) A functional biology of echinoderms. CroomHelm Ltd., London

Lawrence JM, Lawrence AL, Holland ND (1965) Annual cycle inthe size of the gut of the purple sea urchin, Strongylocentrotuspurpuratus (Stimpson). Nature, Lond 205: 12381239

b

Fig. 8 Strongylocentrotus droebachiensis. Indices of food quantity andfood quality (mean SD) at Little Duck Island and Mill Covebetween April (quality) or June 1994 (quantity) and August 1995 inthe kelp bed, grazing front and barren grounds. Food quantity isexpressed as food volume (percentage of total body volume) and foodquality as organic material (percentage of total gut content). Meansare based on 5 to 34 sea urchins

477

Lemire M, Himmelman JH (1996) Relation of food preference tofitness for the green sea urchin, Strongylocentrotus droebac-hiensis. Mar Biol 127: 7378

Levitan DR (1995) The ecology of fertilization in free-spawninginvertebrates. In: McEdward LR (ed) Ecology of marine in-vertebrate larvae. CRC Marine Science Series, CRC Press, Inc.,Boca Raton, pp 123156

Levitan DR, Sewell MA, Chia F-S (1992) How distribution andabundance influence fertilization success in the sea urchinStrongylocentrotus franciscanus. Ecology 73: 248254

Mann KH (1977) Destruction of kelp-beds by sea urchins: a cy-clical phenomenon or irreversible degradation? Helgolanderwiss Meeresunters 30: 455467

Mann KH (1985) Invertebrate behaviour and the structure ofmarine benthic communities. In: Sibley RM, Smith RH (ed)Behavioural ecology. Blackwell Scientific Publications, Oxford,pp 227246

Mann KH, Wright JLC, Welsford BE, Hatfield E (1984) Responsesof the sea urchin Strongylocentrotus droebachiensis (O.F. Mul-ler) to water-borne stimuli from potential predators and po-tential food algae. J exp mar Biol Ecol 79: 233244

Mattison JE, Trent JD, Shanks AL, Akin TB, Pearse JS (1977)Movement and feeding activity of red sea urchins (Strong-ylocentrotus franciscanus) adjacent to a kelp forest. Mar Biol 39:2530

Meidel SK, Scheibling RE (1998) Size and age structure of the seaurchin Strongylocentrotus droebachiensis in dierent habitats.In: Mooi R, Telford M (eds) Echinoderms San Francisco. AABalkema, Rotterdam, pp 737742

Miller RJ (1985) Succession in sea urchin and seaweed abundancein Nova Scotia, Canada. Mar Biol 84: 275286

Miller RJ, Mann KH (1973) Ecological energetics of the seaweedzone in a marine bay on the Atlantic coast of Canada. III.Energy transformations by sea urchins. Mar Biol 18: 99114

Minor MA, Scheibling RE (1997) Eects of food ration andfeeding regime on growth and reproduction of the sea urchinStrongylocentrotus droebachiensis. Mar Biol 129: 159167

Munk JE (1992) Reproduction and growth of green sea urchinsStrongylocentrotus droebachiensis (Muller) near Kodiak, Alas-ka. J Shellfish Res 11: 245254

Pearse JS (1980) Synchronization of gametogenesis in the seaurchins Strongylocentrotus purpuratus and S. franciscanus. In:Clark WH, Adams TS (eds) Advances in invertebrate repro-duction. Elsevier North Holland, Inc., New York, pp 5368

Pennington JT (1985) The ecology of fertilization of echinoid eggs:the consequences of sperm dilution, adult aggregation, andsynchronous spawning. Biol Bull mar biol Lab, Woods Hole169: 417430

Propp MV (1977) Ecology of the sea urchin Strongylocentrotusdroebachiensis of the Barents Sea: metabolism and regulation ofabundance. Soviet J mar Biol 3: 2737

Rogers-Bennett L, Bennett WA, Fastenau HC, Dewees CM (1995)Spatial variation in red sea urchin reproduction and morphol-ogy: implications for harvest refugia. Ecol Applic 5: 11711180

Scheibling RE (1986) Increased macroalgal abundance followingmass mortalities of sea urchins (Strongylocentrotus droebac-hiensis) along the Atlantic coast of Nova Scotia. Oecologia 68:186198

Scheibling RE (1996) The role of predation in regulating sea urchinpopulations in eastern Canada. Oceanol Acta 19: 421430

Scheibling RE, Hennigar AW (1997) Recurrent outbreaks of dis-ease in sea urchins (Strongylocentrotus droebachiensis) in NovaScotia: evidence for a link with large scale meteorologic andoceanographic events. Mar Ecol Prog Ser 152: 155165

Scheibling RE, Hennigar AW, Balch T (1994) The dynamics ofdestructive grazing of kelp beds by sea urchins in Nova Scotia.In: David B, Guille A, Feral J-P, Roux M (eds) Echinodermsthrough time. AA Balkema, Rotterdam, p 871

Scheibling RE, Stephenson RL (1984) Mass mortality of Strong-ylocentrotus droebachiensis (Echinodermata: Echinoidea) oNova Scotia, Canada. Mar Biol 78: 153164

Sivertsen K, Hopkins CCE (1995) Demography of the echinoidStrongylocentrotus droebachiensis related to biotope in northernNorway. In: Skjoldal HR, Hopkins C, Erikstad KE, LeinaasHP (eds) Ecology of fjords and coastal waters. Elsevier ScienceBV, Amsterdam, pp 549571

Sokal RR, Rohlf FJ (1995) Biometry, 3rd edn. WH Freeman andCompany, New York

Starr M, Himmelman JH, Therriault J-C (1990) Direct coupling ofmarine invertebrate spawning with phytoplankton blooms.Science 247: 10711074

Starr M, Himmelman JH, Therriault J-C (1992) Isolation andproperties of a substance from the diatom Phaeodactylum tri-cornutum which induces spawning in the sea urchin Strong-ylocentrotus droebachiensis. Mar Ecol Prog Ser 79: 275287

Starr M, Himmelman JH, Therriault J-C (1993) Environmentalcontrol of green sea urchin, Strongylocentrotus droebachiensis,spawning in the St. Lawrence estuary. Can J Fish aquat Sci-ences 50: 894901

Vadas RL (1977) Preferential feeding: an optimization strategy insea urchins. Ecol Monogr 47: 337371

Wharton WG (1980) The distribution of sea urchindominatedbarren grounds along the south shore of Nova Scotia. In:Pringle JD, Sharp GJ, Caddy JF (eds) Proceedings of theworkshop on the relationship between sea urchin grazing andcommercial plant/animal harvesting. Can tech Rep Fish aquatSciences 954: 3347

Wharton WG, Mann KH (1981) Relationship between destructivegrazing by the sea urchin, Strongylocentrotus droebachiensis,and the abundance of American lobster, Homarus americanus,on the Atlantic coast of Nova Scotia. Can J Fish aquat Sciences38: 13391349

478