Environmental Biology of Fishes

, Volume 91, Issue 3, pp 295–301

Life span, growth and mortality in the western Pacific goby Trimma benjamini, and comparisons with T. nasa


    • Department of Natural HistoryRoyal Ontario Museum
  • Karen M. Alofs
    • Department of Ecology and Evolutionary BiologyUniversity of Toronto
  • Alexandra Marseu
    • Department of Ecology and Evolutionary BiologyUniversity of Toronto

DOI: 10.1007/s10641-011-9782-6

Cite this article as:
Winterbottom, R., Alofs, K.M. & Marseu, A. Environ Biol Fish (2011) 91: 295. doi:10.1007/s10641-011-9782-6


Examination of daily increment rings in the saccular otoliths of 91 specimens of the small goby, Trimma benjamini, reveal a maximum age of 140 days with an average pelagic larval duration of 33.9 ± 4.3 days (SD), or 24.2% of the maximum lifespan. Estimates of daily mortality rate ranged from 2.9% to 6.3%. Comparisons of these results with those for T. nasa suggest that 1) the growth rate of T. benjamini males does not decrease with age as it does for T. nasa; 2) T. benjamini has a longer lifespan and lower daily mortality rate than T. nasa; and 3) T. nasa has a faster growth rate than T. benjamini. These results reinforce the potentially important role of small, planktivorous, outer reef fishes in reef trophodynamics, as well as highlight the need for further research on small reef fishes.


Life historyGobiidaeTrimma benjaminiOtolithsPelagic larval durationAgeGrowthMortalityCryptobenthic fishes


The role of the small cryptobenthic fishes, overlooked for many years mainly because of their low standing biomass, has been the subject of increasing interest to ecologists attempting to understand coral reef trophodynamics (Ackerman and Bellwood 2000; Depczynski et al. 2007). Recent studies have documented that these abundant and speciose taxa represent a significant contribution to both biomass production and energy flow in reef ecosystems (Longenecker and Langston 2005; Depczynski et al. 2007). This information has highlighted the need to consider taxa-specific life history traits when studying reef ecology. Several of these traits, such as mortality, longevity and time of maturation, are strongly correlated with body size and often vary predictably with changes in body size (Depczynski and Bellwood 2006). These features may have a profound effect on the contribution of small fishes to overall reef productivity (Ackerman and Bellwood 2000; Winterbottom and Southcott 2008).

Representatives of the Gobiidae contribute up to 35% of the number of fishes on Indo-Pacific coral reefs (e. g. Ackerman and Bellwood 2000), and up to 20% of the number of species present (RW pers. obs.). Recent studies of the life-history patterns of these small fishes have indicated linear growth trajectories, as well as high mortality and rapid population turnover (e.g. Istigobius decoratus, Kritzer 2002; Eviota sigillata, Depczynski and Bellwood 2005). This suggests that cryptic fish species are ecologically important prey species, and thus make important energetic contributions to the reef ecosystem (Depczynski et al. 2007).

The genus Trimma of the Indo-Pacific coral reefs is composed of colourful, small (mostly <30 mm standard length = SL) gobies and includes 64 valid described species (Winterbottom and Hoese unpubl.), with approximately 30 to 40 other species known but currently undescribed (Saeki et al. 2005; Winterbottom and Southcott 2008). Species for which information is available (mostly unpublished observations by RW) are either sedentary benthic or schooling epibenthic planktivores. Trimma benjamini represents one of the non-schooling, benthic species, and feeds on plankton by making small rapid forays from the substrate into the water column (Saeki et al. 2005; RW pers. obs.).

This study uses data derived from the sagittal otoliths to investigate the age-length relationship, approximate lifespan, time of settlement, and natural mortality rate of Trimma benjamini. Additionally, we compare the life history T. benjamini with that of T. nasa, a closely related schooling, epibenthic species. This paper compliments similar studies on very small (< 3 cm SL) reef fishes such as T. nasa (Winterbottom and Southcott 2008) and three species of Eviota (Depczynski and Bellwood 2006). The information gathered to date will have implications on further studies of reef-trophodynamics, and help to provide a broader understanding of the range of life histories of coral reef fishes (Munday and Jones 1998), their management, and potential conservation strategies.


Fish collection

All specimens were collected with rotenone at Helen Reef (Hotsarihie Reef), Hatohobei State, Republic of Palau, in September, 2008 under a research permit issued by the Palau Bureau of Marine Resources during a survey of the fish fauna of the South West Islands. Lengths of specimens are given as standard length in millimetres. Sixty-eight specimens (10.4–22.9 mm) were collected from the south side of the main channel into the lagoon at about its midpoint at a depth of 8–20 m, and 23 specimens (12.2–19.7 mm) were taken on a vertical outer reef drop-off at the southern end of the reef at a depth of 23–36 m. Specimens were placed in 80% ethanol immediately after collection.

Otolith preparation

Each specimen was measured (SL to the nearest 1/10th mm), sexed (based on the external morphology of the urogential papilla) and then the sagittal otoliths from each side of all specimens were extracted by cutting the head along the mid-sagittal plane and removing the otoliths from the brain cavity. The otoliths were subsequently cleaned in water to remove any adhering tissue, dipped in 75% ethanol and stored dry in plastic culture plates. Thermoplastic glue (Crystalbond 509™) was used to mount each otolith on a glass microscope slide, initially positioned with the sulcus side down and the anterior end, but not the core, overhanging the edge of the slide. Working under a dissecting microscope, 9 and 3 μm lapping film were used to grind the otolith until the anterior end was level with the edge of the slide. The glue was then reheated, and the otolith was repositioned in the middle of the slide with the ground surface down. The posterior end was then ground to obtain a thin section, within which the entire sequence of rings was visible, from the primordium to the outer edge.

Age determination

A Spot Flex model 15.2 camera attached to a Leitz Dialux 22 compound microscope allowed sections to be viewed on a computer monitor using Spot Diagnostic Instruments software (v4.5.9.5). Sections were viewed at 400x magnification. Growth increments were visible and could be counted outwards from the core along the longest axis. Recent validation studies have shown that these increments are deposited daily in many reef-associated, estuarine and temperate marine goby species (Iglesias et al. 1997; Hernaman et al. 2000), as well as in other small reef fishes (e.g. cardinalfishes, unpublished data cited in Longenecker and Langston 2006). We were unable to conduct validation tests on our material, and have assumed that the otolith increments are deposited daily.

The first increment, often called the hatch check, is assumed to be deposited at hatching, mouth opening or absorption of the yolk sac, events that all occur within a 24-hour period in Trimma and allied small gobies (Sunobe 1995; Wittenrich et al. 2007). The presumptive settlement mark, which records the end of the pelagic larval duration (PLD), was also noted. The settlement mark is defined by a characteristic rapid decrease in increment width, with such structural changes apparently coinciding with shifts in habitat (Wilson and McCormick 1999). This mark forms the transition from larval to settled life (see, e.g. Longenecker and Langston 2005, Fig. 1b; Winterbottom and Southcott 2008, Fig. 1). The rings of each otolith were counted twice. A third count was made if the initial counts differed by >10% of the mean. If the two closest then agreed within 10%, their average value was used.
Fig. 1

Trimma benjamini. Age versus standard length with linear (solid) and power (dashed) regression lines. Vertical dotted line indicates the mean age of settlement (33.9 days)

Data analysis

To examine the growth pattern of T. benjamini, linear and power regression lines were fitted to the age-at-length data (Fig. 1). Goodness of fit for each model was determined using the residual sums of squares (RSS), the coefficient of determination (R2) and by examining the distribution of residuals. All mortality was assumed to be natural mortality as T. benjamini is not exploited by humans. Several methods were used to estimate natural mortality. First, the daily mortality rate was determined from the maximum age according to Hoenig’s (1983) equation:
$$ { \ln }M = 1.46 - 1.01{ \ln }T{ \max } $$
where M = instantaneous mortality rate and Tmax = maximum age. Second, we constructed an age frequency histogram of collected individuals (Fig. 2) which was used to estimate survival (S) by Heincke’s (1913) and Robson and Chapman’s (1961) estimators, assuming constant survival and recruitment rates. Instantaneous mortality (M) was calculated from these estimates of S, where S = e-M. Instantaneous mortality also was calculated by fitting a linear regression to the descending limb of the age frequency histogram after frequency was log transformed (Ricker 1975). Instantaneous mortality rates were converted to daily mortality rates for comparison.
Fig. 2

Trimma benjamini age frequency histogram and model used for Ricker approximation of mortality rate

To test for sex-specific differences in T. benjamini growth, a mixed model of the effects of age, sex (male, female or juvenile) and age x sex on standard length was fitted. Data on T. benjamini growth and age at settlement was compared with that of T. nasa, a co-occurring species (data from Winterbottom and Southcott 2008). Age at settlement for the two species was compared with a Wilcoxon two-sample test. A t-test was used to examine differences in the age of sampled adult T. benjamini and T. nasa. A mixed model of the effects of age, species and age x species on standard length was fitted to compare growth rates of the two species. In both mixed models, age was a continuous covariate. Degrees of freedom for comparisons between species were calculated by the Satterthwaite approximation to account for unequal variances between species. Analyses were performed using SAS 9.1 (SAS 2003).


The analyses are based on otoliths from 91 specimens (48 identified as females, 29 males, and 14 immature, 10.4–22.9 mm, with a mean of 17.2). The maximum age recorded was 140 d.

We were able to discern a settlement check mark in all otoliths. The mark was visible as a decrease in increment width, with an accompanying change in increment contrast and focal plane. Settlement mark data indicated the mean pelagic larval duration (PLD) to be 33.9 ± 4.3 days (SD), constituting 24.2% of the maximum recorded lifespan. This corresponds to maximum of 106.1 d of post-settlement life.

The range of planktonic larval duration (PLD) as evidenced by the settlement mark showed much greater variation in T. benjamini than in T. nasa. The range for the former species was 15–45 days (19.3 and 16.2 females respectively) with a mean of 33.9 days, versus 25–43 days (9.3 immature and 15.5 female respectively) with an identical mean value. There was no significant difference between the age at which settlement occurs between the two species (Wilcoxon two-sample test (\( { \Pr } > \left| {\hbox{z}} \right| = 0.5050 \)).

The residuals of linear and power functions were similarly distributed and differed little in their ability to explain variation in increment number (i.e. age) over the range of standard lengths we observed (R2 = 0.6033 and 0.6338, respectively; _ Fig. 1). The similarity between these two models indicated that a linear model appropriately described T. benjamini growth, and growth rate did not appear to change with age.

Using Hoenig’s equation, T. benjamini daily mortality was 2.9%. The age frequency histogram used for further mortality rate estimates used 15 day increment age classes (Fig. 2). Classes with fish younger than 76 days were not included in estimates because their frequency indicates that they may have been undersampled. Daily mortality was 4.0% when calculated using Heincke’s (1913) estimator and 5.6% when calculated using Robson and Chapman’s (1961) estimator. When estimated by linear regression of natural log transformed frequencies (Ricker 1975), the daily mortality rate was 6.3%.

There was a significant difference between species in the age of collected adults (N = 150, Satterthwaite approximation df = 135, t = 15.70, P = 0.0001). In our collections, T. benjamini adults were significantly older (mean = 89.442 days, SD = 15.474) than T. nasa (mean = 55.582 days, SD = 10.605) adults.

The mixed model analysis did not detect any sexual differences in growth (as measured by SL) of T. benjamini (Table 1). Additionally, there was no significant age x sex interaction.
Table 1

Trimma benjamini. Results of mixed model of age, sex, and interaction on standard length (mm)













Age x Sex




The mixed model, including T. benjamini and T. nasa, showed that individual age and species both significantly explained variation in standard length (Table 2). On average, T. benjamini were significantly larger than T. nasa, however, there was also a significant age x species interaction which indicated that T. nasa grows significantly faster than T. benjamini (Fig. 3)
Table 2

Trimma benjamini versus T. nasa. Results of mixed model of age, species and age x species, and interaction on standard length with Satterthwaite approximation of degrees of freedom













Age x Species




Fig. 3

Growth of Trimma benjamini (solid circles and solid line) and T. nasa (open circles and dashed line) based on the results of mixed model


The two species of Trimma compared here are very abundant on the outer reef slopes and drop-offs of many western Pacific reefs, and usually dominate the small reef fish fauna in terms of both numbers and biomass. Trimma benajamini and T. nasa share many life history characteristics common in small reef fish. These characteristics include short lifespan, long PLD relative to lifespan and linear growth rate.

The mean PLD of Trimma benjamini was identical to that of T. nasa (Winterbottom and Southcott 2008) and comparable to that of many coral reef species (Depczynski and Bellwood 2006). More interestingly, this period of time represented 24.2% of the maximum recorded age, a value that is notably higher than the < 1% exhibited by of most other reef fishes (Depczynski and Bellwood 2006). This is congruent with recent data showing that PLD comprises a greater proportion of total lifespan in some small, cryptic species than in other reef fishes. For example, the other gobies Trimma nasa and Eviota sigillata both have a PLD of over 35% of total life expectancy (Depczynski and Bellwood 2006; Winterbottom and Southcott 2008). The apparent fixed range of PLDs for coral reef fishes indicates a constraint which leaves these small short-lived species with relatively little time in which to complete reproduction (Depczynski and Bellwood 2006).

During its rapid post-settlement development, T. benjamini exhibits a linear growth pattern. Long-lived reef fishes often exhibit asymptotic growth patterns which have been attributed to a trade-off between growth and reproduction (Choat and Robertson 2002). In contrast, non-asymptotic or linear growth appears to be common in small reef fishes including both temperate and tropical gobies (Hernaman and Munday 2005). Linear growth in these fishes may be explained by their ability to compensate for the energetic costs of reproduction by shifting to a more profitable or higher quality food source or increasing feeding rate (McCormick 1998).

Sex-based differences in growth rates have been shown in several goby species (Hernaman and Munday 2005). These differences have been attributed to differences in the cost of gamete production or parental care (i.e. egg-guarding) or related to mating systems and sexual selection. For example, Hernaman and Munday (2005) demonstrated more rapid male growth in 2 polygynous goby species but not in 3 monogamous gobies. Sunobe and Nakazono (1990) indicated that Trimma okinawae undergoes bi-directional sex changes which are strongly influenced by body size. The largest female in a population will change sex to become a male when a dominant male is not present, but will revert back to female when a larger male reappears. Winterbottom and Southcott (2008) showed that, after reaching reproductive maturity, growth of male T. nasa slowed while females retained their pre-reproductive growth rates. The mating system of T. benjamini is currently unknown, but based on the specimen analysis certain inferences can be made. For example, the sex-ratio is only slightly-female biased (1:1.6 male/females), which suggests that polygyny is unlikely in this species. There is no evidence of hermaphroditism in this species as evidenced by an intermediate shape to the urogential papilla. Unlike T. nasa, we found no differences in growth rates between the female, male and immature categories of T. benjamini, suggesting that the latter species exhibits the same growth trajectories irrespective of sexual maturity or sex (rather than the ‘broken stick’ model suggested for the former species).

In further constrast with T. nasa, T. benjamini has a longer life-span, lower daily mortality rate, and lower growth rate. The differences in the life history of these closely related species may be related to differences in their habitat use and behavior. Trimma benjamini occurs singly or in loose groups and is benthic, while T. nasa forms schools that hover (usually in a head-up position) close to vertical structures on the drop-off. Miller (1984) showed that nektonic species of temperate gobies and those living in open sandy habitats had shorter life-spans than cryptobenthic and burrow-dwelling species, which presumably had lower predation risks. Hernaman and Munday (2005) suggested that a similar relationship exists between life-span and habitat use in tropical gobies, which is again likely related to predation risk. Estimations of T. benjamini daily mortality rates ranged from 2.9% (Hoening) to 6.3% (Ricker). While the uncertainty associated with these estimates may limit their usefulness in estimating population dynamics, they remain useful for comparisons between species. Daily mortality rates estimated for T. benjamini are lower than those same estimates for T. nasa (4.7% from Hoening’s equation and 8.2% where the Ricker estimate was based on 10-day increments for age classes, excluding fish younger than 51 days).

Trimma benjamini’s lower daily mortality rate and subsequently longer life-span may be related to the lower predation risk associated with its behavior and habitat. Mortality can have important implications in the life history of reef fishes, influencing many growth parameters. For example, rapid growth and maturation likely compensate for increased mortality rates (Reznick and Endler 1982; Depczynski and Bellwood 2005). This may explain why T. nasa exhibits a higher growth rate than T. benjamini.

Despite representing a small part of the static biomass in reef systems, small cryptic fishes are thought to account for a large proportion of overall energy flux due to their high turnover rates (Ackerman and Bellwood 2000; Kritzer 2002; Depczynski et al. 2007). The life histories of these Trimma species indicate they are likely more trophodynamically important than previously recognized. Reef trophodynamics depend, at least in part, on the importation of energy from elsewhere. As planktivores, these two species, which are found mostly on the outer reefs, harvest food from oceanic currents which travel across the reef front carrying allochthanous plankton. The small size, abundance, and high natural mortality rate of these two species suggest that they are important prey species within the reef ecosystem. Their plankton diet and availability as prey for many other reef dwellers suggest their ecological importance as net energy importers to the reef. However, much further research is needed to understand the relationship between the life history characteristics we have described and the ecological roles and trophodynamic importance of these fishes.


Sincerest thanks to the seven other members of the “fish” team that helped to collect the specimens used here, and especially to Mark Westneat (Field Museum, Chicago) and Pat and Lori Colin (Coral Reef Research Centre, Palau), whose financial contributions to the expedition were essential to its realization. Once again, Claire Healy (ROM) generously allowed us almost unlimited access to her compound microscope. Our grateful thanks to Laura Southcott (Department of Zoology, University of British Columbia) for her very useful comments and insights on the draft manuscript. The fieldwork was financially supported by the ROM Foundation, the ROM’s Department of Natural History, NSERC Discovery Grant 7619, an NSERC Ship Time grant, and a grant from The Nature Conservancy (all to RW—my deep gratitude to all these agencies and their officers for making this expedition possible).

Copyright information

© Springer Science+Business Media B.V. 2011