GASTER0STE1TLE. 245 



Leiurus pungitius, Swainson, Fishes, ii, p. 242. 



Gasterostea pungitia, Sauvage, Revis. Epin. 1874, p. 29 ; Moreau, Poiss. France, 

 iii, p. 169. 



Tinker, Couch, Fish. Brit. Isles, i, p. 176, pi. xxxvii, fig. 2. 



American variety. 



Oasterosteus occidentalism Cuv. and Val. iv, p. 509 ; De Kay, New York Fauna, 

 p. 68, pi. xlii, f. 135. 



Gasterosteus pungitius, Storer, Rep. Mass. p. 32. 



Gasterosteus nebulosus, Agassiz, Lake Superior, p. 310, pi. iv, f. 2. 



Pygosteus occidentalis, Brevoort, Amer. Fish. Report, 1873, p. 794. 



Abnormal varieties. 



Gasterosteus pungitins, Day, Jour. Lin. Soc. 1876, Zool. p. 110 {Destitute of 

 ventral fins) . 



B. iii, D. 7-10/^Vo, P- 10, V. 0-1, A. s-Vo, 0. 12. 



Length of head 3| to 4J, of caudal fin 5| to 7j, height of body 5J to 6j in the 

 total length. Eye diameter 3j to 4 in the length of the head : 1| to 2 diameters 

 from the end of the snout, and 3/4 of a diameter apart. This species is more 

 elongated than G. aculeatus. The maxilla extends posteriorly to beneath the 

 front edge of the eye. Teeth villiform in the jaws, more numerous anteriorly 

 than they are laterally, none on palate or tongue. Fins the spines of the first 

 dorsal of varying length, but low, being about equal to half the height of the rays 

 of the soft dorsal, the last of which are the longest. It has been stated that these 

 spines are not in a straight line one behind the other, but that alternate ones are 

 more to the right or left ; this I have not been able to confirm in all the examples 

 which I have examined : their edges are smooth. Caudal square or slightly 

 emarginate at its posterior extremity. Ventrals consisting each of one spine and 

 one ray when present, and which articulate with the pubic plate : an arrow- 

 headed expansion of the humeral arch, which is subject to great variation both as 

 to its width and length, and may even be entirely absent. The sides, as a rule, 

 are entirely smooth, but in some few instances an indistinct caudal keel has been 

 observed. Colours vary considerably, usually of a dark olive-green or brown 

 along the back, becoming lighter on the sides and beneath, with numerous black 

 spots and dots irregularly distributed : fins diaphanous. In the breeding season, 

 those which I have seen become much darker, the under surface of the head and 

 chest turning to a cobalt blue. They are said to assume their nuptial livery 

 about March : Nilsson observes that in Scandinavia they become red at this period. 



Varieties. Inform. Allman obtained in Tipperary several specimens of a 

 stickleback, apparently referrible to G. pungitius, but without any trace of ventral 

 spines (Thompson) : subsequently I captured thirteen examples of this species at 

 Edgeworthstown, all having a badly marked keel on the side of the tail ; eleven 

 of these were destitute of ventral fins. When the spine existed, as a rule it was 

 not half the length of the pubic plate : in one example the spine on one side of 

 the body was one-third shorter than its fellow on the opposite side. In those 

 wherein no ventral spine existed the pubic plate was also absent. The Swedish 

 variety, though generally possessing 9 dorsal spines, has been observed to have 

 them from 8 to 10. Jenyns considered that in the British Isles there are two 

 varieties, the first with the " sides of the tail keeled ; each keel being furnished 

 with several slender scales, which themselves appear scaled under the microscope : 

 secondly G. Icevis, Cuv. and Val. having the sides of the tail quite smooth." 



Donovan observed that we have in some 10 dorsal spines, others with 9, 

 while one specimen had only 8. Shaw also remarks that sometimes, though but 

 rarely, the spines amount to 11. Parnell took two at Prestonpans, one having 9, 

 the other 11 dorsal spines. Among those I personally obtained at Edgeworths- 

 town, and which varied from 1'4 inches to 2T inches in length, one had 2 dorsal 

 spines, only the seventh and ninth being visible : one had 4 spines, and eleven 

 had 8 to 9 spines. Thus we have local instances wherein the dorsal spines 

 numbered 2 > 4, -8,. 9, 10 and 11. 



