ON THE VALUE OK LARVAL CHAKACTERS. 287 



a "low" Saturnian, not a " high " Citheronian ; smiilarly, A(jUa is a 

 " high " Citheronian, not a " low " Saturnian. 



Ail the Emperor moths I have yet been able to examine, allow of 

 their being placed in two family groups, which correspond very nearly 

 to Hiibner's Ecliidnne (my A<iUa<lae), and Heraea (my Saturniadac). 

 Thus, from characters offered by the imago, I offer a classification 

 which allows of a test being made between it and a classification 

 dependent upon the larval tubercles offered by Dr. Dyar. Leaving 

 aside the fact that the morphological value of the larval characters 

 may not be always correctly appreciated by Dr. Dyar, there appears 

 the fact of a bold contradiction between neurational and larval 

 characters, as respectively interpreted. From my point of view, I go 

 as far as this : The acceptance of Dr. Dyar's classification or division 

 of the Emperor moths upon the presence or absence of larval tubercles, 

 as against mine upon the neuration of the perfect insect, involves the 

 rejection of the neuration as of any value whatever. More than this, 

 it involves the rejection of all other characters drawn from the imago, 

 since all other characters fall in with my grouping much more easily 

 and naturally. It is a distinct forcing on general characters to class 

 Hemileuca and Antoynerb (Hi/percliiria) in one "family." No ex- 

 planation is offered, or can be offered, how insects with such contra- 

 dictory types of neuration can be members of the same " family," 

 while other " families " violently claim their interpolation, since they 

 offer precisely the two respective types which are here badly mixed by 

 Dr. Dyar, upon the comparatively unimportant character of the 

 existence or suppression of a larval tubercle. " Stinging spines " are 

 clearly a matter of adaptation, arising independently in different 

 groups. In pointing out elsewhere the value of Dr. Dyar's classifica- 

 tion, arising out of the relative position of the larval tubercles and 

 their arrangement, I have disputed their specialisations as of value to 

 determine the limits of family groups, and for this reason : the 

 specialisations of larvae (and of pup^e and cocoons) follow independent 

 lines, and cannot be brought into coincidence, for the purposes of an 

 artificial system, with the specialisations of the imago. I pointed out 

 this fact as to the larvae more than twenty years ago. The presence, 

 relative condition, or absence of the dorsal tubercle is a character of 

 specialisation. 



Therefore, the issue lies in this : If we are to recognise such larval 

 characters in our system, we must abandon all study of the imago as 

 useless. The Emperor moths are a case in point, and I join issue 

 with Dr. Dyar in the classification of the Satuniiades : and I believe 

 that this matter must be cleared up before we can go any further. 



On the development and probable origin of certain ocellated spots 

 in the larvae of Lasiocampa quercus and Odonestis potatoria. 



By A. BACOT. 



I wish to call attention to the white spots which are present in the 

 centre of the sub-dorsal area of the meso- and post-thoracic segments of 

 the larva of L. quercih. These spots are developed from the ends of 

 the last and largest of the orange transverse bands which cross the 

 dorsal area of these segments, and which, together with similar stripes 



