Systematics of Prickly Sculpin, I — Krejsa 
245 
cies was separable "at least subspecifically from 
the Sacramento River form." Seale (1896) and 
Gilbert (1896) used the valid name for north- 
ern specimens. But, obviously, Jordan (1895) 
and Jordan and Evermann (1896) still thought 
in terms of a distinct Californian species ( gulo - 
sus ) and a distinct northern species ( asper )• 
Jordan misidentified a specimen of asper from 
San Luis Cr., near Avila, Calfornia, as gulo sus. 
Jordan and Evermann listed the range of the 
nominal species from Vancouver Island to 
Oregon, and of gulosus, from California Coast 
Range streams and inland in the San Joaquin R. 
Gilbert (1896) referred to Cottus asper of the 
Columbia and Cottus gulosus of the Sacramento 
as "two species so extremely similar that it is 
difficult to distinguish them.” Jordan and Ever- 
mann (1898) repeated the suggestion of Gil- 
bert and Evermann that the nominal species is 
separable, at least subspecifically, from the 
Sacramento R. form, " Cottus gulosus .” 
That Jordan and Evermann perceived neither 
the conspecific relationship of parvus to asper 
nor the limits of the valid species Cottus gulo- 
sus becomes more fully evident on the next 
page (p. 1945) of their 1898 report. Their 
description of Cottus gulosus (Girard) is taken 
from misidentified specimens of C. asper col- 
lected in San Franciscito Cr., Santa Clara Co., 
California. These were large specimens "3 to 7 
inches in length” and, most significantly, the 
count for anal rays is given as "A. 16 to 18.” 
Both of these characters separate asper from 
gulosus. Furthermore, they include Cottopsis 
parvus Girard, from Monterey, the Presidio, 
Fort Reading, and Petaluma, California, in the 
synonymy of gulosus. 
Snyder (1905) collected and correctly identi- 
fied Cottus asper from the same locality, San 
Franciscito Cr. He was probably the only one 
of his time to understand and explain the true 
relationships of asper, parvus, and gulosus. On 
p. 337, he stated: 
Recent authors have identified the common Sacra- 
mento form which represents the Cottus asper of the 
Columbia River with the Cottopsis gulosus of Girard. 
They have sometimes considered the Sacramento form 
as identical with C. asper and have placed the name 
gulosus in the synonymy of the latter. At other times 
they have considered the species as a slightly differ- 
entiated form worthy of recognition in nomenclature, 
and have used the name gulosus to designate it. The 
former view concerning the species is probably cor- 
rect. The association of the name gulosus with it, 
however, is without warrant. The latter belongs to a 
species easily distinguished from C. asper, differing 
notably in having a much shorter anal fin. There 
are usually fewer dorsal spines and rays, a more 
limited distribution of prickles, and an almost uni- 
form absence of palatine teeth. In C. asper the dor- 
sal has 8 to 10 spines and 19 to 22 articulated rays, 
the anal 16 to 18 rays, while in C. gulosus the dorsal 
has 7 to 9 spines, 17 to 18 rays, the anal 12 to 14 
rays. 
Snyder then continues with a note on habitat 
preference: 
In its distribution C. asper appears to be largely 
confined to the lower courses of the streams, being 
especially abundant near tide water, while C. gulosus 
is found further up, where the water is clear and the 
current rapid. 
Rutter (1908) correctly placed Cottopsis 
parvus into the synonymy of Cottus asper, pre- 
sumably recognizing that the two were con- 
specific. However, he incorrectly synonymized 
Uranidea semiscabra centropleura Eigenmann 
and Eigenmann, which is properly referable to 
Cottus gulosus. 
Snyder was the first to consider a series of 
specimens of the nominal species throughout its 
entire range, as then known, and, in the same 
issue of the Bulletin of the Bureau of Fisheries 
in which Rutter had correctly synonymized 
parvus, he noted the extreme variation of prick- 
ling investment. While recognizing the varia- 
tion between streams, he also noted that the 
prickling variation is common among individ- 
uals from the same stream. 
Although explicitly aware of Snyder’s com- 
ments on asper and gulosus, Evermann and 
Goldsborough (1907^) identified 16 specimens 
of gulosus from Loring and Boca de Quadra, 
Alaska. The reliability of these identifications 
is questionable. Kermode’s listing (1909) of 
Uranidea gulosa from Shawnigan L., Vancouver 
Island is probably a misidentification, since 
Fowler (1923:282) listed the same specimen 
as Cottus asper. 
Snyder (1913, 1916) again recorded the 
occurrence of C. asper and gulosus in differing 
habitats of the same stream. Hubbs (1921) 
recognized Jordan’s earlier misidentification of 
gulosus from San Luis Cr., California. He also 
commented on the variability of prickling in 
C. asper from several streams. 
