CHAEACTERS IN THE TABLES 5 



species always possess tAvo peculiar dark bars on the pleurae, but as some of the 

 members of Culex proper have irregularly marked or spotted pleurae, we have 

 thought it safer not to separate the group on this character, but have included 

 the species with the others. The species of Culex are the most closely allied and 

 the most variable of any of the mosquitoes. We have been obliged to use the 

 ordinary characters of markings and coloration, because other characters do not 

 exist ; but there are none of them that are not subject to variation. For example, 

 the white ring on the proboscis of the female may, in occasional specimens of the 

 same species, be absent; the white rings of the tarsi may, exceptionally, dis- 

 appear ; the markings of the thorax, more especially if pronounced and obvious, 

 may be replaced by a uniform coloration ; the white bands at the bases of the 

 abdominal segments may be absent. We have tried to offset this condition by 

 including some of the more variable species in two or more different places in 

 the table; but, in using it, this matter must not be lost sight of, nor the 

 conclusion hastily reached that a specimen represents a " new " species simply 

 because it can not be determined by the table. The larval table will be found 

 to be more reliable, as the larval characters are less variable than those of the 

 adults ; indeed a knowledge of the larva is essential to a proper comprehension 

 of a species of Culex. 



Before leaving this subject reference should be made to the character of the 

 shape of the proboscis, which we use under dichotomy 45. While often obvious, 

 the difference in certain species is slight, so that it is sometimes difficult to decide 

 whether the proboscis is swollen or not. In such cases we advise the trial of the 

 table under each heading separately. In short, the species of Culex in their 

 colorational characters are illy defined, intergradient, and variable, and if our 

 table is found to be vague and the characters inconstant and unsatisfactory, the 

 fault must be laid at the door of the insects themselves. The variability of the 

 specific characters does not imply that the species are not distinct. Their dis- 

 tinctness is proved by the well-marked differences in the male genitalia, which we 

 find to be reliable, and the same can be said of the larvae. 



The species of Uranotcenia are few in number and beautifully ornamented, 

 lending themselves to specific separation on the usual characters of markings and 

 coloration. The species seem constant and we do not anticipate that there will 

 be any difficulty in recognizing specimens in good condition. Unfortunately 

 the same is not the case with the larvae which present few tangible characters 

 for separation. Mitchell published a table separating species on detailed char- 

 acters of the structures of the antennae, which was followed by Dyar and Knab ; 

 but we now find that these characters are not only beyond the limit of specific 

 fixity, but also beyond the limit of accuracy of observation, so that we have 

 had to discard them entirely. 



The species of Culiseta are well defined and will cause no trouble, once the 

 genus is ascertained. This is in some cases more uncertain than the specific 

 determination, owing to variability in the venation, which affects some of the 

 species more than others. The larvae are more difficult to separate than the 

 adults, although very readily determinable generically. Indeed, we feel some 



