348 BRITISH LEPIDOPTERA. 



appearance at the end of the stage. It is very difficult to see clearly 

 because of the glistening larval surface, and its borders are very ill- 

 defined. It is possible that this feature also belongs to the next 

 stage and is only seen through the skin. Knowing that these larvae 

 subsequently become brown, I was very interested to find that five 

 of them (the first hatched, and four of the fifteen younger larvae) 

 possessed a well-marked brown area, bounded by tubercles bearing 

 bristles, upon the dorsal surface of the prothorax (see fig. 5), another 

 brown triangular area, similarly bounded, on the anal • flap, and a 

 brown quadrangular area on each side of the anus (fig. 6). The anal 

 claspers were also brown. These marks were present through the 

 whole stage ; the posterior patches were of a much darker brown colour 

 than that upon the prothorax. The tubercles and bristles were, of course, 

 similarly arranged in the larvae without the brown areas. [The 

 prothoracic area is, I feel sure, the homologue of the hard plate 

 which occupies this position in wood-boring and burrowing larvae, 

 in Tortricids, &c. This plate is evidently a very ancient feature of 

 the lepidopterous larva. Traces of it can, I believe, be found almost 

 universally at one or other of the larval stages. It will be again 

 referred to in this ontogeny.] These five larvae were isolated in order 

 to determine whether the differences would be increased in succeeding 

 stages. First stadium (mature): About 8*2 5mm. in length when 

 extended although the larvae can stretch to 8 - 6mm. ; their length was very 

 uniform. In the resting-period before ecydsis the larva is cylindrical in 

 shape, long and narrow, with a very uniform diameter, but tapering 

 very slightly and evenly from the 8th abdominal segment to the 

 anterior extremity ; this appearance is especially w r ell seen when 

 the larva is somewhat contracted. Second stadium : Immediately 

 after ecdysis the horn is colourless, but it deepens into black (except 

 for the median zone) in the course of an hour or two, this is also 

 true of the dark patches and shades * present on most of the larvae. 

 The head retains the rounded form, but loses it in subsequent stages. 

 The caudal horn is held as in the last stage, but it is now very nearly 

 straight. It is still bifid, but the fork is much smaller and less 

 conspicuous. Its surface is now thorny, from the presence of 

 relatively few large tubercles which terminate in bristles ; a section 

 of its length nearer the tip than the base is now white and partially 

 transparent. The general appearance of the structure from the 

 anterior side, is seen in pi. xv., fig. 9 ( X 24*5). The larva is still green 

 and is covered with white shagreen-dots, each terminating in a 

 bristle, exactly as in Smerinthus and Sphinx ligustri in this stage. The 

 subdorsal is present, but is not so conspicuous as it becomes later in 

 the stage. All other characters present at the beginning of the stage 

 varied greatly. The main tracheal trunk is quite visible through the 

 whole of this stage, and the dorsal vessel still forms a very prominent 

 dark green median line. On September 20th, when the average length 

 of the seventeen larvae was io^mm., they were all carefully compared 

 together, with the following results : 



* Poullon notes : " This stage opened with a great surprise ; I had fully 

 expected that the live larvae which exhibited dark marks in the last stage, would still 

 continue to be the darkest varieties in this and that the differences would even 

 increase. To my great astonishment these five larvae were much lighter than the 

 others as a whole, so that the relations were completely reversed. Similar reversals 

 occurred in the later stages." 



