C. Information from Samples Taken from Aggregations

Sampling is often an important part of the study of reef fish spawning aggregations. This can include the measuring and weighing of fish, taking whole organs or tissue samples, DNA samples, blood samples, gut contents, and otoliths. Samples provide the data to determine basic information on sex ratios, size frequency distributions, readiness for spawning, and many other facets of the life of the fish. The methods used for sampling an aggregation often determine what information can be reliably determined about the aggregation.

Samples can be obtained by purchasing them from fishers, examining fish at landing points or just prior to sale, or by removing fish from an aggregation using various fishing methods, some of which allow fish to be returned alive. Researchers need to be flexible and opportunistic in acquiring samples, as we are seldom presented with an ideal situation for sampling. In some cases whole fish can be examined in detail if entire fish can be obtained and returned to the laboratory for work-up. Sometimes opportunities may be limited to examining and sampling from dead fish in local fish markets or at sea. Some organs may not be available even if there is a fishery. Ripe ovaries are often considered a particular delicacy, for example, so may not be readily available. Fishers or traders might not want you to handle their fish at all. Also, when looking at fishes in fish markets, always remember they may have been caught in many different places and that the timing of spawning (or other biological factors) can vary somewhat among places, even within the same general region. In some circumstances, fish must be sampled live and returned to the aggregations. There are techniques and sampling considerations for each situation and type of information sought.

Assessing Aggregation Size-Frequency Distributions

Some studies have placed an emphasis on gathering data on the size of fish in an aggregation to follow changes in fish size over time, to determine the proportion of mature size fish present, or for other reasons. Samples from fishermen can prove particularly useful in this regard, either measured at the aggregation as fish are caught, or obtained later from fish markets. It has been common practice to provide size-frequency curves for fish from grouper aggregations (e.g., Carter et al., 1994, Colin, 1992; Colin et al. 1987; Sadovy et al., 1994a). In most cases the sex of the fish is also determined, although this is obviously not possible when captured fish have been gutted. Even gutted fish are of value, however, since their size data can contribute to the overall size frequency profile for an aggregation.

Much care is needed in collecting and then interpreting size-frequency data taken from aggregations. First, it is important that sufficient samples have been taken to ensure a representative number of fish. Very often this will mean that many hundreds of fish need to be measured. Second, it is most important to check whether these samples actually reflect the size distributions of the fish that are present; for example, some fishing methods are size-selective. Ideally, it is preferable to compare sizes assessed underwater (by spearing or by visual estimates) directly with fish sizes taken by fishing gears at the same time and place (e.g. Shapiro et al., 1993).

Interpretation of size-frequency data obtained from spawning aggregations needs care. It is common to find considerable natural variation in strength among year classes in long-lived species. This means that apparent changes in size among a few year classes may not be entirely or largely due to fishing, or, indeed, due to fishing at all. For example, changes in mean length or in the size-frequency distributions from one year to the next could be the result of fishing but they could also arise for other, natural reasons. Many reef fishes are long-lived and can vary considerably in year-class strength. This means that, in any one year, a fishery might be dominated by just one or two age classes (e.g. Russ et al., 1996). Any inter-annual changes in size distributions noted from size estimates, therefore, could be due to natural causes, to fishing induced causes, or, of course, to some combination of the two. As an example, differences were noted in the size frequency distributions of aggregated fishes taken over six consecutive years from the same aggregation site of red hind grouper sampled in the same way each year, with the same fishers, and the fish lengths determined with a high degree of accuracy in the laboratory (Fig. 20). In this particular case, how do we know whether the size differences over the six-year period are due to natural variability in recruitment and inter-annual growth, or to impacts of fishing (compare 1987 with 1992 for example and see the recruitment suggested in 1989)? Moreover, if time-series of size frequency data are too short, relative to the longevity of the study species, then it may not be possible to distinguish changes in the shape of

Figure 20. Size-frequency distribution for red hind, Epinephelus guttatus, taken by hook and line fishing, and by the same fishers, on spawning aggregations at Bajo el Cico, Puerto Rico from 1987 to 1992. n=number of fish taken each year. (Figure redrawn from Sadovy et al., 1994a, with kind permission of Kluwer Academic Publishers).

size frequency distributions over time due to fishing from changes due to major recruitment differences among years. Taking the argument one step further, if changes in size frequency distribution are noted and they can somehow be demonstrated to be due to the impacts of fishing, how do you know if they are due to aggregation or non-aggregation-based fishing? These are important questions to those of us who work on aggregations, and seek to propose or support their management using such information. The examples serve to emphasize the need for understanding both the limits of methodology (in this case body size estimation) as well as the biology and fishery of the study species to ensure meaningful and useful interpretations of data to be made.

Methods for Length and Weight Determination

In the field it is most common to simply take the total length of the fish and its weight. When you are working on a fishing vessel or in a fish market, speed is often times very important, particularly if there are 100 or so fish waiting to be measured, and the fisher wants to get them cleaned and on ice. In such cases, it is best to have two people involved; one to handle and measure the fish, a second to record data. If additional data and samples are being taken, such as gonads, even more people would probably be beneficial.

In most cases the total length (TL) of the fish will be determined as it is simple to do with a measuring board. A measuring board has a meter stick inset into a plywood block so it is flush, and a board sticking up at a right angle along one side. The fish is put onto the board, the snout placed against the upright where the meter stick starts, and the length read off the meter stick at the end of the tail. It takes about 10-15 seconds to measure a fish using this board.

If necessary, a simple tape measure can be used, however remember to take into account the curvature of the tape when laid along a large fish. Weight can be taken with a hanging spring balance of the type used by most sport fishermen for weighing their catch. However, such scales are often inaccurate and need to be carefully calibrated. A more expensive hanging pan balance with a dial indicator is better, but again should be carefully calibrated both before and after weighing. If you need to check calibration in the field, but lack standard weights to do so, take a measuring cup with which you can accurately measure out 1 liter of water (or any multiple thereof); this amount of water can be put in a plastic bag, and weighed using the scale. It should weigh approximately 1 kg per liter. Saltwater and the plastic bag add a few grams of extra weight, but are insignificant against the 1 kg of water given the precision of the equipment used.

Length and weight data should ideally be recorded on prepared data sheets, with additional spaces for sex, gonad condition and other information. If time allows, it might be useful to measure both total length (TL) and standard length (SL), particularly for those species where a TL-SL conversion might not be easily available. Standard length is the length from the tip of the snout to the end of the vertebral column. The end of the vertebral column can usually be determined by bending the tail of the fish so a distinct crease becomes visible at the end of the spinal column. For species with a forked tail, such as jacks and trevalleys (Carangidae) the fork length (FL) can be measured: from the tip of the snout to the center of the tail. For many species with a deeply forked tail, SL is difficult to determine as the caudal peduncle does not bend easily.

Figure 21. Size frequency of Nassau groupers, Epinephelus striatus, caught by hook and line fishers, Long Island, Bahamas. Unsexed fish had been gutted prior to measuring, but still provide useful information on the overall size of aggregating fish. The small size of some males implied that, perhaps, some males were not from sex-changed females; an observation which was further pursued. After Colin, 1996.

In nearly all cases, size frequency data from collected samples should be considered superior to visual length estimates taken underwater (Fig. 21). However, there is ample opportunity for bias in the part of the aggregation that ends up captured by fishermen. At least, if samples are collected in a consistent manner over a number of years, say from fish caught by fishermen using the same methods each year, they should be comparable and provide firm data on relative changes in fish size over time. A comparison between visual estimates and caught fish would be interesting for a single aggregation. For returning fish alive, see section on handling live fish below.