DIMORPHIC QUEENS IN AN AMERICAN ANT. 159 



that the -female must be sterile, notwithstanding her well-de- 

 veloped ovaries, or the two species would long since have merged 

 into one. 1 



(c] It seems improbable that such an aberrant creature as the 

 /?-female would mate with the male of another species, but this 

 argument loses much of its force when we stop to reflect that the 

 claviger male is very similar to the latipcs male even in the struc- 

 ture of its genitalia. 2 



4. We may suppose that we are dealing with a true case of 

 dimorphism in the female sex. On first thought this seems im- 

 probable because dimorphic queens, in the strict sense of the 

 term, are unknown among ants. But when we stop to consider 

 that the social bees and wasps exhibit an essentially similar dimor- 

 phism, except that one of the two winged forms, the worker, is 

 sterile (and this may also be the case with the a-female of L. 

 latipcs /) there is nothing preposterous in this view. Moreover, 

 in ants the wingless workers have themselves in many species 

 become dimorphic, developing soldier and typical worker forms, 

 either perfectly distinct from each other or connected by a series 

 of intermediates. Why, then, may we not expect the winged 

 queens in some cases to exhibit dimorphism among themselves, 

 especially when dimorphism " runs in the blood," so to speak, of 

 all the social Hymenoptera ? And why may not L. latipcs be 

 such a species in which the old and deeply-rooted tendency is 

 breaking out in a novel form ? This would at least complete the 

 theoretical possibilities in female ants as represented in the fol- 

 fowing diagram : 



Primitive female. 



Worker, or sterile female. Queen, or fertile female. 



Worker minor. Worker major, or soldier. a-female. 3-female. 



It thus appears that of the four hypotheses, two may be rejected 

 as too improbable to be entertained, and that the true meaning of 



1 Unless, indeed, the Mendelian law be supposed to operate with unprecedented 

 clearness in this particular case. 



2 In this connection, however, it is interesting to note that Marchal ('96, p. 45) 

 failed to induce mating between two very closely allied species of wasp ( I'espa get - 

 nianicii and I', vnl^aris). 



