1887.] between certain Pupce and their Surroundings. 99 



if corroborated, may be physiologically connected with pigment 

 formation, or it may merely give the larva an additional opportunity 

 of being acted on by light, if for any cause the illumination of the 

 surrounding surface is delayed, or if the most sensitive part of the 

 whole period corresponds to the ordinary darkness of night. 



6. Experiments which show the Sensitive Condition during Stages (ii) 

 and (iii). — It was very important to obtain beyond any doubt the 

 demonstration that the larvae are sensitive during Stage (ii), and also 

 to decide conclusively whether any susceptibility was continued into 

 Stage (iii), and if so to compare the relative susceptibilities of the two 

 stages. Such experiments, if successful, would at once dispose of the 

 older theory of pupal sensitiveness, and would be most important in 

 making possible other methods of investigation which, applied to 

 Stage (iii) alone, might successfully terminate the long and difficult 

 search for the larval sensory surface which is affected by surrounding 

 colours. A great many experiments were conducted with this object. 

 The larvae were made to pass Stages (i) and (ii) exposed to the influence 

 of a powerfully acting colour, and then were transferred for Stage (iii) 

 to the colour which tended most strongly in the opposite direction. 

 The largest, most carefully conducted, and most successful experiment 

 of the kind gave the following results, all the larvae belonging to the 

 same company : — 



Degrees of colour. 



(1.) 



(2.) 



Dark 

 (3.) 



(3.) 



Light 

 (3.) 



(4.) 



(5.) 





In black surroundings for the 























1 





5 





1 





= 7 



Transferred from black into 



















gold for Stage (iii) 









1 



5 



3 





= 9 



Transferred from gold into 



















black for Stage (iii) 











6 



9 





= 15 



In gold surroundings for the 



















wbole period 











5 



7 



8 



= 20 



















51 



The analysis speaks for itself. Stages (ii) and (iii) are both sensitive, 

 but Stage (iii) is much less sensitive than the other. Thus, when the 

 earlier part of the period was passed in gilt surroundings, the resem- 

 blance between the results and those produced by gilt surroundings 

 acting during the whole period was much stronger than the resem- 

 blance between the latter and the results produced when the gilt 

 acted during Stage (iii) only. It is observable that the larvae as a 

 whole evidently tended towards the lighter forms, so that the black 



