inai-kiiuis mul uttittidcs of Icpidoptcrom hirvcc. 303 



extremity (page 528), Finally, in certain individuals of 

 M. stellatarum, I found distinct traces of the fork in the 

 last stage. Comparing these observations, the structure 

 appears to be most persistent in the larva {M. stellatarum) 

 with a primitive form of marking, longitudinal stripes ; 

 and especially prominent in early stages of another 

 larva with the same marking [A. jnnastri). I have no 

 doubt that it is largely developed in the first stage of 

 M. stcllatamm also. It further occurs in the earlier 

 stages of the larvae with the more advanced oblique 

 stripes at the time when the primitive markings (after- 

 wards lost or much diminished) are well developed upon 

 them. This is especially well seen in S. In/ustri, where 

 the fork is very distinct in the first stage, when the sub- 

 dorsal is also the most prominent marking, while the 

 fork ceases to be recognisable about the time when the 

 last traces of the subdorsal disappear (after the third 

 ecdysis). AVe are thus led to the conclusion that the 

 forking of the horn is a primitive character, of historic 

 value only in the ontogeny, and remaining longest in 

 forms that have other primitive features persistent. The 

 horn was also covered with hairs (in common with the 

 rest of the body), each of which projected from a 

 tubercle. The terminal prong consisted of two enlarged 

 diverging tubercles terminated by large hairs or bristles. 

 In the difticult question of assigning a function to this 

 structure the bifid termination must be taken into con- 

 sideration. So also the immense size of the primitive 

 horn must be rememl)ered, shown by its greater relative 

 predominance in the earlier stages of the ontogeny, 

 especially in the case of S. Ihjmtn. Finally, great 

 importance must be attached to the fact that the horn 

 is movable in the two first stages of S. Ufitistri, and 

 entirely under the control of the animal's will. These 

 facts seem to indicate that the horn was primarily a 

 defensive structure. Further, Mr. Meldola, in the above- 

 mentioned Appendix (page 527), states that the caudal 

 horn of Ouerocampa Lycetiis is freely movable, and he 

 suggests that the horn may be " a remnant of a fiagellate 

 organ having a similar function to the head-tentacles 

 of the i%jj//io-larva\ or to the caudal appendages 

 of Dicranura.'' A curious point was incidentally dis- 

 covered in examining the larvae \mder a compound 

 microscope. The caudal horn exhibits distinct move- 

 ments synchronous with the contractions of the dorsal 



