![Thumbnail: [Volume 84, 1956-57 page 312]](/journals/rsnz/images/rsnz_84/thumbnails/rsnz_84_02_0359_0312_ac_02.gif)
Laboratory Technique
For the purposes of sorting and counting, samples were divided into two categories; those with less than 200 (approximately) and those with more than 200 spherical teleostean eggs. For the smaller samples a complete count was made. The entire sample was placed in a rectangular counting dish, 50 × 30 mm and 5 mm deep, the bottom of which was ruled lengthwise with numbered parallel lines 2 mm apart (i.e., slightly less than the width of the field of a 16 mm microscope objective) Using a mechanical stage, it was then possible to scan the entire sample line by line, identifying and counting snapper eggs. (Other species were also counted, but the results are not relevant to this discussion.)
The principal character employed in identification was the diameter of the egg capsule, which was determined with an eyepiece micrometer. Very few eggs were found in the range 0.81 to 1.02 mm which showed any characteristics other than those of snapper. It is, of course, not impossible that some of the eggs identified as snapper could have belonged to other species. After weighing all available evidence (Cassie, 1956) it has been concluded that this is unlikely, but until the eggs of all species spawning in the Gulf are known there can be no absolute certainty in identification. The frequency distribution of measurements tended to have two or (occasionally) three or more modes. Since the size of eggs taken from individual fish had been found to be approximately normally distributed, it was suspected that two populations (or even two species) of eggs might be present. However, on analysis of a number of such distributions on probability paper (Cassie 1954) it was found that the two apparent size groups could not be matched from sample to sample, so that it is doubtful whether any significance can be attributed to them. In the preserved state, eggs could be divided into two distinct categories according to whether the yolk sac was clear or opaque However, it was found that artificially fertilised eggs from a single fish also exhibited this differentiation, which is therefore assumed to be an artifact.
The larger samples, sometimes containing several thousand eggs, could not conveniently be counted directly in this manner, and for these a sub-sampling apparatus as shown in Fig. 3 was constructed. The sample container, A, is a 250 millilitre beaker, and the stirrer, B, is an annulus of 16 gauge brass, 5.6 cms outside and 3 cms inside diameter attached to a brass handle. A straight glass sub-sampling tube, C, 2.7 mm in diameter and 100 cms long, has a scale, D, calibrated in ml near its upper extremity. The sub-sampling tube communicates through a stop-cock, E, with a mercury aspirator consisting of two glass tubes, F and H, 13 mm in diameter and
![Thumbnail: [Volume 84, 1956-57 page 313]](/journals/rsnz/images/rsnz_84/thumbnails/rsnz_84_02_0360_0313_ac_02.gif)
30 cms long, connected by a rubber pressure tube, G. The top of the open arm has a constriction, K, which damps the oscillation of the mercury. The open arm is attached to a slide, L, which can be raised through a distance of 20 cms and fastened in the upper position by the peg, M.
It was originally hoped that direct counts of snapper eggs could be made while they were in the sampling tube. This was eventually found to be impracticable, since snapper eggs could not always be reliably distinguished by the unaided eye from those of other species of similar size. However, it was found possible to make counts of all eggs (disregarding species) with considerable speed and accuracy, so that a convenient test of the validity of the sub-sampling technique could be made by this method. This test will be described in the next section.
Fig. 3.—Apparatus for sub-sampling fish eggs from plankton samples.
The detailed procedure for operating the sub-sampling apparatus was varied as more experience was gained, but the following scheme was ultimately developed as standard, although in some cases it was slightly modified to cope with the variable nature of the samples. The sample is diluted with tap water to a volume of 100 ml
![Thumbnail: [Volume 84, 1956-57 page 314]](/journals/rsnz/images/rsnz_84/thumbnails/rsnz_84_02_0361_0314_ac_02.gif)
and poured into the beaker, B, which is raised on a wooden block so that the end of the sub-sampling tube is immersed. The tap, E, is then opened and the slide raised to its upper position. The tap is closed and the slide lowered, thereby creating a negative pressure between the tap and the right-hand column of mercury. The sample is agitated by moving the stirrer rapidly up and down, and at the same time the tap is opened, allowing a sub-sample to be drawn into the sub-sampling tube. The volume of water was then recorded from the scale and the sub-sample allowed to drain into the counting dish. An egg count was then made in the same manner as for the smaller samples. Each sample was sub-sampled a sufficient number of times to give a minimum count of 200 snapper eggs. The total number, N, of snapper eggs in the sample was then estimated as:
N = 100 n/v
where n = total snapper egg count, all sub-samples.
v = total volume of all sub-samples (ml)
It will be shown in the next section that, although the volume of sub-samples has a slight variation, little error would be introduced by taking it as constant and employing a mean value of v in the above calculation.
Before using the sub-sampling apparatus it was often necessary to remove larger plankton which would not pass into the sampling tube. Whenever possible this was accomplished by washing through a brass wire gauze sieve with approximately 4 meshes to the centimetre. Organisms which could not be removed in this way were picked out by hand. Copepods and other small animals were left in the sample as they could be easily distinguished from eggs when counting. Some difficulty was encountered when Thallia was present, since eggs tended to adhere to the salps and thus could not readily be separated. However, since the tows were of shorter duration, such samples rarely contained large numbers of eggs, and it was possible to make a direct count, removing eggs individually with a pipette.