Sparks : Icn thyological Survey ahout San Juan Islands. 195 



it is fruin S 109.5.'' Tlu' Cdiulilion nf tlic prcxciit ral plales, which may 

 usually !)(.' depended upon to separate tliese two species, is not abso- 

 luteK- dei)en{hU)le, as occasionally A', penlacanthus has only one pair, 

 as in A'. iiiJrdspiiKiliis, instead of two, its normal number. 



A', iiifraspiiiatns resembles A', aluskauus much more closely, but 

 may at once be known by the well develoiK'd spines on the lower lateral 

 bod\' ridge. These are sharp, stand well awa>- from the body, and do 

 not decrease in size to the base of the caudal fin. In A", alaskaniis 

 these sjiines are reduced to very small needle-like points lying close 

 to the bi)(l\, and on the caudal ju'duncle are scarcely distinguishable 

 by the naked eye, though they may be readily felt by the finger. 



The body of these specimens is scarcely slenderer than that of X. 

 alaskaniis, nor are the spines and ridges on the side of the head weaker. 

 The postocuiar spines are usually smaller, and the postocular and 

 nuchal depressions wider and deeper. The anal opening though very 

 often farther back than it e\er is in X. alaskaniis, is, on the other hand, 

 frequently as far for\vard as in that species. This variation is not 

 se.xual. 



The small point extending forward on the rostral ])late noticed 

 by Dr. Gilbert (Proc. U. S. Nat. Mus., XXVUI, p. 982) is often absent 

 especially in large specimens. In the same paper a lapsus calami may 

 be here corrected. In stating the width of the interorbital as being 

 equal to three-fourths of the diameter of the orbit, the intention was 

 to give three-fourths of the diameter of the pupil. In the specimens 

 at hand this dimension varies from three-fourths to the full diameter 

 of the pupil. The variation is that of the pupil rather than that of 

 the interorbital. 



46. Averruncus emmelane Jordan and Starks. 



This was found to be the commonest agonoid fish among the islands. 

 Specimens were taken in abundance in the trawl at from fifteen to 

 forty fathoms. 



Of the sixteen specimens examined, ten have nine dorsal spines, and 

 six have eight; tw^elve have eight dorsal rays, and four have seven; 

 eleven have eleven anal rays, and fi\e have twelve. The pores of 



< The greater variation in A', penlacanthus goes with the greater discrepancy in 

 size of the specimens measured in that species. They were from three and a half 

 to seven inches long, the larger ones the slenderer. In X. infraspinatus the speci- 

 mens were from three to four and one-half inches. 



