significantly fewer gill rakers than a sympatric population of normal size in Trout Lake, 

 Yukon Territory. These examples demonstrate that dwarfing need not be accomplished by 

 planktivory or by gill raker increase. 



We conclude that the morphological, ecological and reproductive data indicate that the 

 pygmy smelt is a distinct species from the rainbow smelt. 



We consider the pygmy smelt as a sibling species because no single character 

 distinguishes it from the rainbow smelt and multivariate analysis is needed to identify all 

 non-spawners. In the usage of Rising and Schueler (1980) it might therefore be called a 

 morphological sibling species. However, those authors suggest calling siblings those 

 species having more within-group variation than between-group variation. But in our case, 

 the among-group variation, as measured by a one way ANOVA of the first principal component, 

 was 91.6% and exceeded the within-group variation, 8.4%. However, these percentages might 

 be expected to change if populations of pygmy smelt are analyzed and when anadromous rainbow 

 smelt samples are added. 



Nomenclature 



Copeman (1977) has shown that anadromous and lacustrine rainbow smelt are distinguished 

 from one another at a level approximately characteristic of subspecies. For the present, 

 because lacustrine rainbow smelt, occurring in Newfoundland and on the mainland, do not have 

 a continuous distribution required of a subspecies (Bailey, Winn and Smith, 1954), 

 lacustrine and anadromous rainbow smelt will be treated for the moment as a monotypic 

 species. The name Osmerus mordax (Mitchill, 1815) is the earliest name available for this 

 form; Klyukanov (1969) refers the eastern North American smelt to 0. m. mordax (Mitchill, 

 1815), distinguishing it from £. m. dentex Steindachner , 1870 of the Pacific and Arctic 

 oceans and drainages. 



Concerning the pygmy smelt, only two lacustrine populations of smelt have been named in 

 eastern North America, 0. spectrum in Wilton Pond, Maine and 0. abbottii in Cobessicontic 

 Lake, Maine, both by Cope (1870). The type specimens of both are missing, as noted earlier. 

 Topotypic material of Osmerus abbottii is not available to us and Cope's original 

 description does not permit us to determine which of the two species of smelts he had in 

 hand. 



Topotypic material of Osmerus spectrum and the small size of Cope's specimens (89 mm SL) , 

 said to be breeding, suggest that he most likely had the pygmy smelt in hand. But Cope's 

 description when compared with that of Delisle (1969a) leave some doubt as to the identity 

 of Cope's specimens. While the length of the head, 4.25 times in length, eye one third of 

 head length, and the 10 anal fin rays favour identification with the pygmy smelt, the high 

 number of lateral line scales is characteristic of the rainbow smelt. It is possible that 

 both pygmy and rainbow smelt are found in the lake. Cope did mention that larger specimens 

 were known and exhibited the characteristic features of the species. Kendall (1927) 

 reported a specimen 5-7/8 inches (133 mm SL) long. Because there are some legitimate doubts 

 as to the identity of Cope's types and because his type specimens are missing, it appears 

 desirable to fix the identity of Cope's name. We therefore designate NMC79-0834, a specimen 



23 



