Text-fig. 1—Relationship between the number of dorsal and the number of ventral teeth on the rostrum of Palaemon affinis. Fig. 1.—50 specimens from throughout New Zealand. Fig. 2.—50 specimens from Island Bay, Wellington. Fig. 3.—50 specimens from Point Howard, Wellington.

Variation in the rostral formula.

Scatter diagrams showing the different rostral formulae and the numbers of specimens having each formula, have been constructed (Text-fig. 1) for three samples of Palaemon affinis. In selecting the material for analysis, obviously damaged specimens and those showing malformed or recently damaged and regenerating rostra (e.g., Text-fig. 2, Figs. 15–21) were rejected. These will be discussed later. Fig. 1 shows the variation in sample 1, a random collection of 50 specimens as follows:

Waiheke Island, Auckland (10); Ohope Beach, Bay of Plenty (3); Wellington Harbour (27); Portobello, Dunedin (7); Ocean Bay and estuary of River Nairn, Chatham Islands (3).

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

A small amount of variation is represented, only 7 different formulae being present. All these specimens would fit into a general formula of 7 to 9/3 to 5, while 50% have the formula 8/4. The Wellington Harbour specimens (just over 50% of the sample) were neither from Island Bay nor Pt. Howard and were in the restricted range 8/3 and 8 to 9/4 with 50% having the formula 8/4.

Fig. 2 shows the rostral variation in sample 2, of 50 specimens from one intertidal pool at Island Bay, Wellington, collected on 3rd October, 1953. The scatter shows a much wider range of variation than that in Fig. 1. Not only are there more formulae.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

represented, but the sample is more evenly spread among several different formulae. In this sample all the specimens would fit into the general formula 5 to 9/2 to 4 while the formulae 6/3, 7/3 and 8/4 each comprise about 20% of this sample.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Fig. 3 shows the variation in sample 3, also of 50 specimens from one intertidal pool at Pt. Howard, inside Wellington Harbour, and approximately 8 miles from Island Bay, which is on the coast outside the harbour. This sample collected on 21st November, 1953, shows more variation than sample 1, but not as much as in the Island Bay sample. All specimens fit into the formula 7 to 10/3 to 5 while almost 50% have the formula 8/4. Sample 3 has a very similar frequency distribution to that of sample 1, but both are quite different to sample 2, where almost 50% have the formulae 6 or 7/3 which were found on less than 5% of each of the other two samples.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Therefore a general formula covering the whole 150 specimens represented in the graphs would be 5 to 10/2 to 2 with the commonest individual formula being 8/4 (about 40% of the 150 specimens). This variation in the rostral formula does not bear any relation to sex. The general formula given above has no specific value, it probably does not establish the full range of variation in this species, and no doubt further collecting would reveal several more unusual formulae. It is, however, useful to give an indication of the variation to be expected in the rostrum of Palaemon affinis and probably in any other closely allied species.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

It is interesting to compare here Thomson (1903). He “examined an immense number (of P. affinis) from various parts of the Colony” and found the variation to be 3 to 6/7 to 9–and states that he “never, however, found a specimen with only six.

Text-fig. 2.—Examples of different rostra of Palaemon affinis. All to same scale, setae omitted. Fig. 3.—Waiheke Island, Auckland. Figs. 1, 2, 4–6, 8, 9, 12, Island Bay, Wellington. Figs. 7, 13, 20, Lyall Bay, Wellington. Figs. 10, 11, 17, 18, 21, Pt. Howard, Wellington Figs. 15, 16, 19, Wellington. Fig. 14,—Portobello, Dunedin.

teeth on the upper margin of the rostrum”. Thomson's papers show that it was his practice to include carapace teeth in the rostral count. Fig. 2 shows that 25% of the Island Bay sample had six dorsal teeth.

Variation in the shape of the rostrum.

The length of the rostrum when graphed against the greatest depth, in sample 2, showed only a small amount of variation. There was no significant change in proportion in these dimensions, the variation at any one size being small, and there was no sexual differentiation in the sample.

The profile of the rostrum varied least of all throughout the entire collection. With the exception of the three forms described below, all others showed the slightly upturned profile described and illustrated by Yaldwyn (1954) and illustrated here in Figs. 1, 3, 6–8 and 10 on Text-fig. 2. The other three forms seen are illustrated in Figs. 9, 13 and 14. Fig. 9 is a down-curved rostrum which was seen only once in this study, Fig. 13 is an upturned sabre-shaped form and Fig. 14 is a long, straight, attenuated form. Both the latter appear to be uncommon, since in the 300 specimens examined each was seen on not more than three specimens.

Ratio of rostral length to carapace length.

The length of the rostrum, measured as c in Text-fig. 3, Fig. 1, was compared with the length of the carapace, measured as d, for samples 1 and 2. The rostrum increased in length in proportion to the length of the carapare, but while these proportions were variable, variation was within narrow limits. For sample 1, the rostrum expressed as parts per 100 of the carapace varied from 77 to 140 and for sample 2 from 82 to 130 (see Text-fig. 5, Fig. 1). In this sample 32% of the specimens had the rostrum shorter than the carapace, 60% had the rostrum longer, while the remaining 8% had the rostrum and carapace subequal. Thus the length of the rostrum was sometimes shorter but more often a little longer than that of the carapace.

Position of 1st dorsal tooth.

The position of the anterior tooth of the dorsal series is very variable. Milne-Edwards (1837) in describing Palaemon affinis from New Zealand implies that the extremity of the rostrum is bifid. But when describing P. quoianus, also from New Zealand (1837), he writes, “Rostre droit, robuste,.… armé de six dents en dessus et de trois en dessous, et point bifide à l'extrémité, mais terminé par une seule pointe, à la base de laquelle sont placées, immédiatement au-dessus l'une de l'autre de la première dent de la rangée supérieure et celle de la rangée inférieure”. Milne-Edwards therefore separates two forms of Palaemon from New Zealand, one with a bifid tip to the rostrum and one without. It has long been recognised that P. quoianus is a synonym of P. affinis. Miers (1876) states that “the number of teeth varies slightly in a large series of specimens, while the bifid appearance is caused by the greater or less approximation of the anterior tooth of the upper series to the apex of the rostrum… also a variable character”. The present material bears this statement out completely. All intermediate stages between the almost bifid appearance of what must be taken as typical P. affinis and the single acute point of typical “P. quoianus” have been seen. Text-fig. 2, Figs. 1–12, illustrate the varying position of the 1st dorsal tooth. These camera lucida outlines include the dorsal aspect of the carapace. They are taken from adult P. affinis with rostra showing no signs of damage or malformation and represent different rostral armament. It is not known if regeneration ultimately reproduces the original formula of a specimen. The figures are arranged in the plate in the following order:—Figs. 1—5 have 3 ventral teeth: Figs. 6—10 have 4 ventral teeth; Figs. 1 and 6 have the 1st dorsal tooth so near the tip of the rostrum that it appears bifid; Fig. 11 is similar, but has 5 ventral teeth; Figs. 2 and 7 have the 1st dorsal tooth immediately subterminal, thus the tip of the