122 



HARDWICKE'S SCIENCE-GOSSIP. 



to be got. Undoubtedly bees would not continue to 

 visit flowers simply because they possess attractive 

 colours. They require, as Sir John Lubbock points 

 out, much more substantial inducements. 



Secondly, with regard to Mr. Bulman's attack on 

 the theory, that special colours are developed by 

 special insects ; red and blue by bees, for instance. 



In the first place, the experiments of Sir John 

 Lubbock which Mr. Bulman mentions, were not 

 intended to prove that theory at all. In Sir John's 

 notice of them in his " British Wild Flowers in 

 relation to Insects" (p. 12), he does not even 

 mention red or blue as the colours chosen. The 

 experiments were of course intended to show that 

 bees can remember, and therefore distinguish, in- 

 dividual colours. These experiments are recorded at 

 length in "Ants, Bees, and Wasps" {International 

 Scientific Series), pp. 291-302. Another set of 

 experiments, recorded in the same volume, pp. 303- 

 307, do certainly show the preference of bees for 

 honey placed on blue paper ; and although Mr. 

 Bulman considers it "more philosophical and con- 

 clusive " to study the habits of bees with regard to 

 flowers themselves, I would ask why bees so greatly 

 prefer honey on blue paper, if not because they prefer 

 the colour blue in flowers ? 



At any rate, whether bees prefer blue and red to 

 other colours or not, I suppose no one has ever 

 asserted that they do so in an resthetic sense. The 

 contention of the upholders of this theory is rather 

 that they have learned to consider blue or red as an 

 index of high specialisation ; and therefore of flowers, 

 which, while presenting peculiar adaptations for their 

 visits, in many cases exclude (by the length of their 

 tubes, etc.) other insects, except of course Lepi- 

 doptera ; that red and blue are in fact highly-evolved 

 colours, consequently present in highly-evolved 

 flowers which are fitted for the visits of highly- 

 evolved insects. 



To make this point more clear, it may be as well 

 here to mention some important conclusions of 

 Hermann Muller, and Grant Allen. Almost the 

 last words of Midler's great work, " Die Befruchtung 

 der Blumen," are : — 



" On the whole we find red, violet, and blue colours 

 appearing for the first time in flowers whose honey is 

 quite concealed and which are visited by more or less 

 long-tongued insects (bees, long-tongued flies, Lepi- 

 doptera), or else in flowers visited for the sake of 

 their pollen chiefly by bees and drone-flies {Hepatica 

 triloba, Verbascum phaniceum)." 



Grant Allen shows a remarkable accordance with 

 this view, which is all the more striking because 

 his work is on such entirely different lines ; his object 

 being to account for the colours of flowers, while 

 Miiller's is to explain their mechanisms in relation to 

 their insect visitors. In summing up his " Law of 

 Progressive Colouration," he comes to six conclusions ; 

 the three bearing on this point are : — 



(1). The most advanced members of all families 

 are usually red, purple, or blue. 



(2). Almost all the members of the most advanced 

 families are purple or blue. 



(3). The most advanced members of the most 

 advanced families are almost always blue, unless 

 spotted or variegated. 



N.B. The qualifying words, "usually," "almost," 

 are accounted for by his theory of " retrogression." 



Now Grant Allen here designates by the word 

 "advanced" precisely those species which have 

 become so specialised, that their honey is inaccessible 

 to any insects except bees, Lepidoptera, etc. 



Grant Allen's theory of colour development, as 

 expressed in his " Colours of Flowers," can hardly 

 be accepted in its entirety. Much of it undoubtedly 

 requires further proof, and much is unlikely on 

 general scientific grounds ; but however much or 

 little of it we accept, his accordance with Muller on 

 these points is worth noticing. 



As to Mr. Bulman's first objection to this theory, 

 viz. that certain blue and red flowers are seldom or 

 never visited by bees, I do not think the facts support 

 him. 



Personally, I have not systematically observed the 

 visits of insects to flowers ; but his observations on 

 two or three of the red and blue flowers mentioned 

 by him as seldom or never visited by bees are not 

 supported by those of Hermann Muller, recorded in 

 his "Befruchtung der Blumen," of the patience and 

 care of which I need hardly speak. These two or 

 three (the common poppy and the periwinkles) are 

 the only native German flowers (except the Scillas) 

 in Mr. Bulman's list, and therefore the only ones 

 recorded in Miiller's work. 



Of these, the common poppy {Papaver Rhceas) has 

 been observed by Muller to be visited by seven species 

 of bees, and only three species of all other kinds of 

 insects. 



Vinca minor has a similar record, viz. bees, seven ; 

 other kinds of insects, three. 



On Vinca major Muller has observed only Bombus 

 agrornm. 



Mr. Bulman's second ground of objection to this 

 theory is that bees often visit flowers of other colours 

 than red or blue. This no one will deny. Indeed 

 Hermann Muller in his " Befruchtung " comes to the 

 conclusion, that " the study of particular species of 

 insects confirms the conclusion based on observation 

 of the more conspicuous flowers, that in general 

 anthophilous insects are not confined by hereditary 

 instinct to certain flowers, but fly about seeking their 

 food on whatever flowers they can find it." 



I think, in fact, we may safely conclude that while 

 red and blue, appearing as they do in flowers highly 

 developed in other respects, were evolved through 

 the selective action of long-tongued insects such as 

 bees and diurnal Lepidoptera, by which they are in 

 many cases almost exclusively visited, on the other 



