Natural history of Emerita analoga ( Stimpson ) ( Anomura , Hippidae ) in a sandy beach of northern Chile Historia natural de Emerita analoga ( Stimpson )

Se realizaron muestreos de sedimentos en la playa de Hornitos, norte de Chile (ca. 22°26´S) durante Agosto y Noviembre de 1996 y Marzo, Julio y Noviembre de 1997, con el objetivo de estudiar la biologia reproductiva y poblacional de Emerita analoga (Stimpson) (Decapoda, Anomura) en una region de la costa Chilena, caracterizada por la presencia de surgencias persistentes. Las mayores abundancias de E. analoga ocurrieron durante Marzo y Diciembre de 1997. Durante Noviembre de 1996, Marzo y Diciembre de 1997, las abundancias mas altas ocurrieron bajo la linea de efluente, mientras que durante Julio de 1996 y Agosto de 1997, la mayoria de las estadios E. analoga ocurrieron con mayor abundancia sobre esa linea. Se recolectaron hembras ovigeras durante todo el periodo de estudio. Los juveniles tambien estuvieron presentes durante todo el muestreo, registrandose las mayores abundancias durante Marzo de 1997. La talla de las hembras ovigeras vario entre 12,6 y 26,4 mm. La fecundidad aumento significativamente con la talla de las hembras en cada periodo estacional analizado. El analisis del patron de reclutamiento de machos y hembras sugiere la existencia de dos pulsos en el reclutamiento de E. analoga en Hornitos: uno entre Octubre y Diciembre, y otro entre Febrero y Marzo. Los analisis del crecimiento muestran que las hembras tienen tasas de crecimiento y tamanos corporales maximos mayores que los machos. Comparaciones con estudios similares realizados en la costa del centro sur de Chile (ca. 39° S), muestran similares tasas de crecimiento y tallas corporales maximas de E. analoga mayores en el sur que en la poblacion de Hornitos. Por lo tanto, se concluye que la cercania a areas de surgencias persistentes parece no afectar la historia de vida de E. analoga en el norte de Chile


INTRODUCTION
Exposed sandy beaches are the most common coastal habitats along the coast of Chile, north of Isla de Chiloe (ea.42° S).In terms of species richness, the intertidal macroinfauna living in those habitats is characterized by dominance of crustaceans, mainly peracarids (J aramillo 1987, 1994).In terms of abundance, the anomuran crab Emerita analog a (Stimpson) is one of the dominant species along all this coast and contributes more biomass by far than any other species (Jaramillo et al. 1993(Jaramillo et al. , 1998)), a similar situation to its northern hemisphere counterpart inhabiting sandy beaches of California (Dugan & Hubbard 1996).
In Chile, the reproductive and population biology of Emerita analoga has been studied in sandy beaches of central (Osorio et al. 1967, Conan et al. 1975) and south central Chile (Contreras et al. 1999).So far, no studies have been carried out on sandy beaches of northern Chile, a coastal area characterized by persistent upwelling events (Fonseca & Farfas 1987, Marfn et al. 1993).In contrast, upwelling only occasionally influences coastal areas located further south.It has been argued that primary productivity is enhanced by upwelling waters, which in turn increase secondary production of benthic invertebrates (Tarazona et al. 1988, Arntz et al. 1987, Bosman et al. 1987).Bearing in mind that upwelling waters result in higher primary productivity at the coast of northern Chile (Marfn  et al. 1993, Rodriguez 1998 1 ) and that temperature affects biology of invertebrates (Brusca & Wallerstein 1979), it is reasonable to expect that northern populations of Emerita analoga will differ in their reproductive and life history characteristics from populations located further south on the Chilean coast.To evaluate this, we carried out seasonal samplings at a sandy beach located close to the upwelling center of Mejillones (ea.22° S) and compared the results with that already published for the coast of south central Chile (Contreras et al. 1999).

Study area and physical measurements
The study beach was located at Hornitos (22° 26' S, 70° 13' W), on the coast of Antofagasta, northern Chile.The length of the beach is approximately 4000 m and is backed by a cliff rising to about 50 m above sea level.
The morphology of the sampling site (i.e., beach face slope at the site of the transects) was determined by Emery's profiling technique (Emery 1961).Sediment samples for grain size analysis were collected from the uppermost layer ( 40 mm) of sediment at each macroinfaunal sampling station (see below) with a 35 mm diameter plastic cylinder.The analyses were carried out using a settling tube (Emery 1938).Mean grain size was calculated according to the moments computational method (Seward -Thompson & Hails 1973) and used to estimate sand fall velocity according to the method described by Gibbs et al. (1971 ).Temporal variability of grain sizes was analyzed with ANOV A (Sokal & Rohlf 1995).
Wave height was estimated by measuring the height of breaking waves (n = 1 0) with graduated poles against the horizon, and adding the result to the height difference between the location of the observer and the lowest point where the backwash met the next incoming swash bore.The wave period (measured with a stop watch) was the time interval between breaking waves.Temporal variability of that wave characteristics was analyzed with ANOV A (Sokal & Rohlf 1995).From estimated mean wave height, wave period and sand fall velocity, Dean's parameter (Q) (Short & Wright 1983) was calculated.Surf water temperature was measured at each sampling period with a mercury thermometer (0.1 o C of precision).Water samples collected from the surf zone were used for salinity determinations in the laboratory.

Sampling of Emerita analoga and data analyses
Seasonal samples were collected during spring low tides of July and November 1996 and March, August and December 1997.The specimens of Emerita analoga used for this study came from a general intertidal macroinfauna survey.In that survey, sediments were collected with plastic cylinders (250 mm in diameter) at ten equally spaced levels (stations) along four replicated transects ( separated by 1 m) extending from above the drift line to the swash zone.The uppermost station of each transect was located above the drift line, the second on the drift line and the last one at the seaward limit of the swash zone (indicated by bore collapse).The sediment was sieved through 1 mm mesh and the specimens were stored in 5 % formalin until sorting.Abundances per m 2 were used to draw kite diagrams to analyze zonation of the different population stages.Individuals were measured using the length of the cephalothorax (thus, Le= body size) i.e., from the tip of the rostrum to the distal scoop of the cephalothorax.The smallest individuals (i.e., Le < 5 mm) were measured using a dissection glass equipped with a graticule, whilst the largest individuals were measured with 0.01 mm precision calipers.Animals were classified as males, females without eggs (hereafter females), ovigerous females and juveniles (i.e.individuals smaller than 4 mm).Sexes were distinguished on the basis of anatomical studies on species of the genus Emerita carried out by Knox & Boolotian (1963), Osorio et al. (1967) andPenchaszadeh ( 1971).Temporal variability of the abundances of the total population and that of each life stage was analyzed with ANOVA on log (n+1) transformed data and the a posteriori LSD (least significant difference) test (Sokal & Rohlf 1995).
Fecundity analyses were carried out for each sampling period, except July 1996 when no ovigerous females were collected.The egg mass was carefully extracted from each female, and the diameters of 10 eggs were measured to estimate mean volume (using the formula for the volume of a sphere).The total number of eggs in each mass was determined by displacement of distilled water in a graduated test tube (0.1 ml accuracy) (Diaz et al. 1983).Fecundity values were log transformed to calculate regressions between number of eggs carried by ovigerous females against body size.Differences of slopes and adjusted averages among regression lines estimated for each season were analyzed with ANCOV A (Sokal & Rohlf 1995).
For the determination of growth parameters, the ELEFAN (Electronic Length Frequency Analyses; see Gayanilo et al. 1989) routine of the FISAT Program (Gayanilo et al. 1996) was applied to the seasonal composition of the population by sizes, discriminated by sex.The method assumes that body growth follows the von Bertalanffy Growth Equation (seasonally oscillating version) (von Bertalanffy 1938), (Hoenig and Hanumara, 1982;Sommers, 1988).For this analysis, class intervals of 2 mm for males and females were chosen following the criteria of Wolff ( 1989) and Sparre (1989).Longevity was estimated on the basis of the maximum observed length and the growth parameters derived from length-frequency distribution analyses.
The ELEFAN program does not assume normality in the length frequency distributions, but uses the information on the height and shape of the age distributions identified in the length frequency data.It identifies peaks and troughs in the samples and then fits the growth curve which passes trough a maximum number of peaks (Pauly et al. 1984 ).An index of goodnes of fit, called Rn, is determined by the exponential form of the ESP I ASP ratio (Gayanilo et al. 1989).
To compare different growth rate estimates, the standard growth index lj>' (phi prime: Pauly & Munro 1984, Vakily 1990) was employed as a measure of overall growth performance (Sparre et al. 1989).This rationale provides a unified parameter of growth performance which does not show variations as large as K and Loo values (Sparre et al. 1989).
The recruitment pattern was analyzed through the estimated growth parameters (discriminated by sex) by projecting the length -frequency data backward onto the time axis (Pauly et al. 1984, Gayanilo et al. 1989).Normal distribution patterns in recruitment were distinguished using the maximum likelihood approach «NORMSEP -Hasselblad» contained in the FISAT Program (see Gayanilo et al., 1996 for further details).

Surf and beach characteristics
Water temperature in the surf zone of the beach at Hornitos showed a seasonal trend, with the highest values during March and December 1997 (22 and 24 o C, respectively) and the lowest during August 1997 (15 o C).Water salinity showed little variation, with a range of; 32.2 -33.6 %o.
During the the sampling period the wave heights varied between 1188 and 1562 mm, while wave period varied between 12.6 and 16.6 seconds (Table 1).Results of ANOVA and LSD test show that waves were significantly higher during November 1996 and August 1997.No significant differences were found when seasonal wave periods were compared.Mean grain sizes generally corresponded to fine sands (Folk 1980), with a significant increase in grain size during August and December 1997 (means close to 250 microns) (Table 1).The following values of Dean's parameter (Q) were estimated: July 1996 = 2.8, November 1996 = 2.3, March 1997 = 3.6, August 1997 = 3.8 and December 1997 = 2.2.
According to these values, the beach at Hornitos can be classified as low intermediate (values of Q between 1 and 6, sensu Short & Wright 1983).The largest change in beach face slope occurred between the first three samplings.The beach was steepest during July 1-996 (1/19); between this date and November 1996, the beach slope flattened to 1/29.By March 1997 the slope had steepened once again to 1124, followed by 1/23 and 1/25 during August and December 1997, respectively.

Abundance and intertidal distribution of Emerita analog a
The total abundance of E. analoga peaked significantly during March and December 1997 (65682 and 159354 ind m• 1 , respectively) (Table 2).Juveniles, males and females peaked significantly at the same months, while ovigerous females had their maximum abundances during March and August 1997 (Table 2).During November 1996, March and December 1997, the highest abundances occurred below the effluent line (water table outcrop), while during July 1996 and August 1997, most of the stages of E. analog a occurred in higher abundances above the effluent line (Fig. 1).In general, all stages occupied the same beach levels.However ovigerous females tended to occur lower down in the swash zone than other stages in July 1996 and December 1997 (Fig. 1).

Reproductive biology and fecundity
Ovigerous females were recorded over the entire study period.However, the greatest abundances were registered during the winter sampling of 1997 (August) (Table 2).Juveniles were also present during all sampling months, peaking during March 1997.
Body size of ovigerous females varied between 12.6 and 26.4 mm Le.Fecundity varied between 1386 and 11937 eggs per ovigerous female (Table 3 ).Fecundity increased significantly with size of females at each seasonal period analyzed (Table 3, Fig. 2).The estimated fecundity of females with a body size of 20 mm Le showed its highest value during spring (November) and the lowest at the end of summer (March) (Table 3).The results of ANCOV A analyses showed significant differences between slopes and adjusted means of the seasonal regression lines (Table 4) indicating that the relationship between number of eggs and body size of ovigerous females varied between sampling periods.

Recruitment and growth
As mentioned above, juveniles were collected throughout the study period; they peaked during March 1997 (about 59% of the whole population).The highest frequencies of small males and females occurred during March and December 1997, when they represented a percentage close to 70% (Fig. 3).The analyses of recruitment patterns for males and females (based on FISA T program) suggest the existence of two pulses in the recruitment of E. analoga at Hornitos.Thus, the bimodal shape ofrecruitment (Fig. 4) could be attributed to the presence of a recruitment peak between October and December, and another one around March (cf.Fig. 3).
Growth parameters of E. analoga at Hornitos show that females have higher growth rates and greater maximum body sizes than males (see values of Loo and growth ind~x <1>' at Table 5).
Values of C and Wp (Winter point) suggest seasonal growth oscillations and slowdown of growth during the winter (cf.Table 5 and Fig. 3).The maximum longevity for individuals both sexes was estimated as 3.5 years.

DISCUSSION
The results of this study showed both similarities and differences with previous studies carried out along the distributional range of E. analoga on the Pacific coast between Chile and California.As in the present study, it has been reported that E. analoga reproduces all year round in sandy   Zonaci6n intermareal de Ios diferentes estadios de vida de E. analoga en la playa de Hornitos.Se presentan Ios perfiles de play a para cada fecha de muestreo.La areas negras corresponden a las zonas de resaca y resurgencia; Jas flechas indican la posici6n de la lfnea de efluente.Comparaciones entre el numero de huevos y el tamafio corporal de hembras ovfgeras de E. analoga en cada perfodo estacional.
beaches of central Chile (Osorio et al. 1967), south central Chile (Contreras et al. 1999) and California (Cox & Dudley 1968).With respect to recruitment, however, Osorio et al. (1967) andEfford ( 1965), found just one recruitment period occurring during spring in a sandy beach of central Chile (El Tabo, ea.33°S) and California (ea.33°N).However, Conan et al. (1975) found two Ranges in body size and eggs number carried by ovigerous females of Emerita analoga, results of regression analyses between body size and number of eggs, number of data (n) and estimated fecundity for an ovigerous female with Le = 20.0 mm Rangos del tamafio corporal y numero de huevos portados por Jas hembras de Emerita analoga, resultados de Ios analisis de regresi6n entre tamafio corporal y numero de huevos, numero de datos y fecundidad estimada para una hembra de Le = recruitment periods (during autumn and springsummer) at another site in central Chile (Valparafso, ea.32°S), a situation similar to that found by Contreras et al. ( 1999) further south (ea.39°S).These differences may well be related to recruitment failures reported for coastal benthic invertebrates with planktonic larval stages during sporadic oceanographic events such as ENSO (cf.Moreno et al. 1998).Due to the fact that the larvae of E. analoga remain in the plankton for a relatively long time (3-4 months, Johnson 1939), it can be argued that changes in water conditions will affect them, and consequently benthic populations will also be affected.It is also possible that differences in analytical methods could result in the observed differences since both the present study and that of Contreras et al. ( 1999) used sofisticated statistical packages (i.e., FISA T, Gayanilo et al 1996) for analysis of recruitment patterns, which were not available at the time Osorio et al. ( 1967) and Efford ( 1965) carried out their studies.
During the winter samplings of July 1996 and August 1997, most of the population of E. analog a was found above the effluent line.That situation did not show any relationship with the seasonal variability of the beach characteristics studied; i.e. wave characteristics and mean grain size did not TABLES Growth parameters of Emerita analoga at the beaches of Hornitos and Mehufn (data from Contreras et al. 1999)  differ significantly during that two winter samplings (see Table 1).Since this species is tidal migrant (MacGinitie 1938, Fusaro 1980), it may be argued that the higher shore distribution of E. analoga during July 1996 and August 1997, as compared to the other months, is the result of tidal variability.Thus, even when this species move~ up and down the beach following the upper limit of the swash, there are some times (e.g.different tidal ranges) when the migrant population may be left behind by the outgoing tide forming patches of stranded animals.
Intermediate sandy beaches (such as Hornitos) are the most variable along a gradient of morphodynamic beach types (Short & Wright 1983).Since physical variability in wave energy, grain size, beach face slope and topographic features result in different swash climates (McArdle & McLachlan 1994), it may be hypothesized that intermediate beaches will have more variable swashes which will affect the migration of E. analoga and thus its intertidal zonation.
It has been predicted that marine invertebrates living in colder waters or higher latitudes present larger body sizes and lower growth rates than ...   , 1936).Even though the studies of reproductive biology of E. analog a are scarce along the Chilean coast and carried during different seasons, it is possible to make some comparisons.For example, we estimated a body size of 12-13 mm as the minimum size at sexual maturity for crabs collected at Hornitos, whilst Osorio et al. (1967) reported a minimum of 15-16 mm for specimens collected at El Tabo (about 800 km south of Hornitos).Furthermore, Contreras et al. (1999) found that that minimum size at maturity increased to about 16-18 mm at a beach located in south central Chile (Mehufn, about 2000 km south of Hornitos).Previous studies showed that maximum sizes (represented by the parameter Loo) are higher in higher latitudes (Templeman 1936, Abele 1982).Thus, while in this study we estimated a maximum body size of 18 mm and 28 for males and females collected at Hornitos, Contreras et al (1999) registered values between 23-26 mm for males and 33-34 mm for females (Table 5).
Although Osorio et al. (1967) and Conan et al. (1975) do not give values of Loo, the maximum length of males was close to 19 mm and that of females about 29 mm.These latitudinal gradients in body sizes of E. analoga are similars to those found by Dugan et al. (1991) for Californian populations.On the other hand, growth rates (as shown by the values of parameter cp') of males and females at Hornitos were similar to those of southern populations (see Table 5), although the ranges in that parameter were higher at the northern population.
The growth curves reported for E. analoga in the present study as well as thus reported by Contreras et al (1999), show a seasonal pattern of growth.Although Osorio et al. (1967) and Conan et al. (1975) do not report growth parameters, both agree that during winter time a slowdown of growth occurs.Seasonality is an important characteristic in growth of short lived animals (Pauly 1990, Sparre, 1990) and is usually the result of environmental variability, primarily oscillations of water temperature between winter and summer (Pauly, 1985).Similarly, Fusaro (1978) reported differences in growth rates of two populations occurring in Californian beaches with differing temperatures.
As mentioned earlier, the beach at Hornitos is close to the upwelling center of Mejillones, a situation which could result in faster growth rates fuelled by enhancement of primary productivity at these northern area of the Chilean coast (cf.Marin et al. 1993).However, no such situation was found in this study, again demonstrating that proximity to a persistent upwelling area does not necessarily result in higher growth rates or larger body sizes of sandy beach invertebrates (cf.Jaramillo et al. 1998).The latitudinal differences observed in reproductive and population biology of E. analoga along the Chilean coast show that studies on basic biology of benthic invertebrate species showing an ample distribution along this coast should include latitudinal comparisons (cf. Avellanal et al. 2000).This factor may play a significant role in conservation of benthic communities and the formulation management plans.If, for example, latitudinal comparisons identify differing patterns in biological parameters, differing conservation or management strategies may be required.

ACKNOWLEDGMENTS
To Marcelo Oliva and Ruben Escribano (Universidad de Antofagasta) for logistic support in Antofagasta.To the following students for their help during field samplings: Manuel Rojo, Mauricio Bringas, Hector Castro, Luis Ossand6n, Alvaro Quintana, Luis Vergara and Monique Sanchez (Universidad de Antofagasta).To Marcia Gonzalez (Universidad Austral de Chile) for laboratory analyses.This study was financed by CONICYT (Proyecto FONDECYT 950001) and the final data analyses and manuscript writing benefitted from the financial support given to EJ and PQ by Proyecto FONDAP 0 & BM (Programa Mayor no 3).
Fig.l.Intertidal zonation of the different life stages of E. analoga at the beach of Homitos.Beach face slopes are presented for each sampling date.The black areas correspond to the swash and resurgence zones; the arrows indicate the position of the effluent line.

Fig. 2 .
Fig.2.Comparisons between the number of eggs and body size of ovigerous females of E. analoga for each seasonal period.

Fig. 3 .
Fig. 3. Size frequency distribution and growth curves of females and males of E. analoga estimated by FISA T Program.Distribuci6n de frecuencias de tallas y curvas de crecimiento de hembras y machos de E. analoga estimados a partir del Programa FISA T.

Fig. 4 .
Fig. 4. Recruitment pattern of females and males of E. analoga estimated by FISAT Program.The bell shaped curves indicate the number of recruitment pulses throughout the study period.

TABLE I
Results of the ANOV A and the a posteriori test LSD on the mean of wave heigth, wave period and grain size during the study period.Means linked by lines do not differ significantly Resultados -de Ios ANDEV A y del test a posteriori LSD de Ios promedios de la altura de !as olas, periodo y tamafio medio del grano, durante el periodo de estudio.Los promedios unidos por lineas no difieren significativamente

TABLE2
Results of the ANOV A and the a posteriori test LSD on the mean abundance (ind m -1) of the total population, juveniles, males, females and ovigerous females of E. analoga.Mean abundances linked by a lines do not differ significantly Resultados de Ios ANDEV A y del test a posteriori LSD de Ios datos de abundancia promedio del total de la poblaci6n, juveniles, machos, hembras y hembras ovigeras de E. analoga.Las abundancias promedio unidas por lfneas no difieren significativamente

TABLE 3
Results of AN COV A analyses carried out to compare seasonal regression lines describing the relationships between body size and number of eggs in ovigerous females of Emerita analoga.Seasonal sampling (months) linked by a line do not differ significantly Resultados de Ios AN COV A realizados para comparar !as lfneas estacionales de regresi6n que describen !as relaciones entre tamaiio corporal y numero de huevos en hembras ovfgeras de Emerita analoga.Los muestreos estacionales (meses) unidos por lfneas no difieren significativamente estimated by FISAT Program I = goodness of fit index (explained sum of peaks I available sum of peaks) I 10 (seeGayanilo et al. 1989)