Curtis: Validation of a method for estimating annual fecundity for Hippocampus guttulatus 



333 



latter inference is supported by close correspondence 

 between observed male and female reproductive 

 activity observed over time on the grid and the 

 values predicted by the ISR model (Fig. 3). 



Discussion 



This article presents the first estimates of realized 

 annual fecundity in a population of wild European 

 long-snouted seahorses (H. guttulatus) and vali- 

 dates a potentially nondestructive application of 

 the spawning fraction method (Hunter and Leong, 

 1981) for estimating spawning frequency. Results 

 indicate that estimates of spawning frequency and 

 annual fecundity based on egg production alone may 

 lead to significant overestimates of realized annual 

 fecundity and thus underscore the importance of 

 testing assumptions when using proxies for estimat- 

 ing fecundity in multiple spawning fishes. 



Batch fecundity 



A significant relationship between male size and 

 brood size of//, guttulatus means that size-specific 

 realized annual fecundities can be calculated and 

 converted to age-specific fecundities (e.g., Curtis, 

 2004) by using a length-at-age model developed for 

 H. guttulatus (Curtis and Vincent, 2006). Among 

 fishes, fecundity is strongly size-dependent (Davis 

 and West, 1993; Lowerre-Barbieri et al., 1996). 

 Female size is typically used to predict clutch size in 

 fishes (Bagenal, 1978), including syngnathids (Teix- 

 eira and Musick, 2001; Vincent and Giles, 2003), but 

 male dimensions can be also be used to predict brood 

 size in fishes that provide paternal care, such as syn- 

 gnathids (Strawn, 1958; Teixeira and Musick, 2001) 

 and mouthbrooding cardinalfishes (Okuda et al., 

 1998; Kolm, 2002). Because both the volume of the 

 female's abdominal cavity and volume of the male's 

 sealed brood pouch potentially limit the number or 

 size (or both) of embryos that can be successfully 

 produced by seahorses (Boisseau, 1967), correlations 

 between the dimensions of both parents and brood 

 size likely reflect mutual mate selection for size (Vin- 

 cent and Sadler, 1995; Teixeira and Musick, 2001; 

 Vincent and Giles, 2003). A positive correlation between 

 the standard lengths of//, guttulatus males and females 

 engaged in courtship behavior on the grid (J. Curtis, 

 unpubl. data), indicates that mating is size-assortative 

 in this population. 



Annual spawning frequency 



The underwater visual census data indicate that on 

 average, female H. guttulatus prepared significantly 

 more clutches of eggs (1.2-1.7 times as many) than males 

 brooded in the Ria Formosa in 2001 and 2002 (Table 2). 

 This means that estimates of spawning frequency and 

 annual fecundity based on female egg production (e.g.. 



1.00 



^ 0.75 - 



^ 0.50 



0.25 - 



o 



c 

 o 



o 



Q. 

 O 



0.00 

 0.4 



0.3 



0.2 



0.1 



0.0 



J FMAM J JASOND 



B 



y = 0.S5e 



■1^)1 



V V 



Figure 3 



Within-year variation in the reproductive activity of Hip- 

 pocampus guttulatus (long-snouted seahorse) in the Ria 

 Formosa: (A) the proportion of males with full pouches, and 

 (B) the proportion of females preparing eggs for mating. 

 The curves (equations given) were fitted to catch data from 

 2000 (solid squares), 2001 (solid circles), and 2002 (solid 

 triangles). The reproductive activity of males (open, upward- 

 pointing triangles) and females (open, downward-pointing 

 triangles) observed during underwater visual censuses of 

 the grid (2001-2002) are plotted against census date for 

 comparison. 



histological assessments) may significantly overestimate 

 the actual number of young produced. Brood production 

 by males was probably not limited by the availability of 

 mature females because sex ratios were slightly biased 

 in favor of females (Curtis and Vincent, 2006), and the 

 interclutch intervals of females were equal to the brood- 

 ing periods of males. Significantly longer interbrood 

 intervals for the population of male H. guttulatus in the 

 present study may have derived in part from depreda- 

 tion by other fishes during egg transfer to male brood 

 pouches. This depredation occurred during one of three 

 matings witnessed on the grid. Approximately 1% of 

 male H. guttulatus in the Ria Formosa had holes in 

 their pouch. Such injuries indicate that other predators 



