418 Mr. F. Merrifield and Mr. E. B. Poulton on 



an investigation. A few such experiments upon P. raj^ie 

 were attempted by G. C. Griffiths in 1887 (Trans. Ent. 

 Soc. 1888, pp. 265, 266), and the results upon the whole 

 supported those obtained in the case of V. urtiae. But 

 the experiments were not very convincing because the 

 colours employed were not those which produce the most 

 marked and opposite effects. 



In 1888 I made some experiments of the kind upon 

 P. rcqim and P. hrassiciB (Trans. Ent. Soc. 1892, pp. 445, 

 446 and 484), using a box lined with black and orange 

 S((uares, but owing to the excessive mortality from the 

 attacks of ichneumons only 6 of the latter and 2 of the 

 former could be tabulated. The results liowever entirely 

 confirmed the experiments made in 1886 upon V. urticie, 

 intermediate and not parti-coloured pupse being always 

 obtained. In 1892 I made a large number of experiments 

 upon Vanessa io (I. c. pp. 420 — 426), and again upon 

 V.itrticie {I. c. pp.391 — ^397). Furthermore a method had 

 by then been arrived at which enabled the larvse of the 

 Vanessidm to be subjected to conflicting colours during 

 the Avhole of the sensitive period ; and another method 

 whereby the dorsal and ventral surfaces could be subjected 

 to opposing stimuli (in the case of V. io, I. c). In all 

 previous exjDeriments the anterior and posterior parts of the 

 body had been thus treated. All such modifications and 

 additions yielded confirmatory results. 



I was nevertheless very anxious to repeat the experi- 

 ments upon the Pierinm and therefore took advantage of 

 the abundance of P. hrassicm last year (1898). Here, 

 again, however, owing to the Ichncumonidm, my results 

 were not at all what I had anticipated ; but taking them 

 in combination with those wl.ich have been already 

 published they leave little or no room for doubt. 



Two conflicting colour-boxes were made for me by Mr. 

 W. Holland and Mr. A. H. Hamm. The first box had an 

 internal section of 54*5 cm. by 14"2 cm., and a depth of 

 7'4 cm. (from back to front in the position made use of). 

 It was used resting on one long side with a clear glass 

 front. The roof (the side uppermost in the position in 

 which a receptacle is used is here always called the " roof") 

 was divided into 9 bays of about equal size by means 

 of 8 hanging partitions (each 7'4 cm. deep and thus 

 extending from the back to the clear glass front, and 

 hanging down for a distance of 3*5 cm.) ; while the back 



