Nov. 20, 1873] 



NA TURE 



45 



end, and prolonged above into orange-coloured trian- 

 gular appendages of their connectives (c, Figs. 21, 22), 

 and lie so close together round the style, as to 

 form a hollow cone containing the pollen, and over- 

 topped only by the skull-like crest of the style. This 

 position of the stigmatic knob rising out of the anther- 

 cone but immediately below its summit, is secured 

 by a remarkable contrivance, the skull-like knob being 

 prevented from sliding into the anther-cone by two tufts 

 of hairs, projecting like whiskers from its two cheek-like 

 lateral surfaces. Thus a lifting up of the stigmatic knob, 

 which must always be effected by insects seeking for honey 

 or for pollen, and which is easily accomplished by them 

 in consequence of the base of the style being slender and 

 bent like a knee, will be more likely to tear off the filaments 

 than to push the stigmatic knob into the anther-cone. 

 Indeed, we find that by the swelling of the fertilised ovary 

 the filaments are always torn off, whereas the anthers re- 

 main, enclosing like a hollow cone the narrow portion of 

 the style, and the skuU-hke knob is never drawn between 

 the anthers. If the anther-cone containing the pollen 

 were densely closed all round, the pollen-grains would not 

 fall out unless the anthers were separated from each 

 other by lifting up the stigmatic knob ; but there actually 

 exists an opening on the lower side of the summit of the 

 cone directed downwards, the appendages of the two lower 

 anthers being cut out [op. Figs. 21, 22), by which nearly 

 all the pollen may fall out spontaneously. When it has 

 fallen out, a great part of the pollen is collected in the 

 close hairy lining of the fore part of the spur. 



Thus far the two forms of Viola tricolor are iden- 

 tical in structure ; and the same, or nearly the same, in- 

 sects may a priori be supposed and have really been 

 observed, to visit the two forms. The distance between 

 the closed entrance of the flower and the honey con- 

 tained in the uppermost part of its spur being in both 

 of the two forms 6-7 mm., an insect must be provided, 

 in order to reach the honey, with a proboscis of at least 

 that length, imless it be enabled by its small size to crawl 

 with its whole body into the flower. A proboscis of 6-7 

 mm. length or larger is only to be met with among all our 

 insects in Lepidoptera, Apidas, and some few Diptera ; 

 insects sufficienily minute to be able to crawl into and 

 out of the flowers, are to be found chiefly in the genera 

 Thrips and MeHgethes. It may therefore be supposed, a 

 priori, that Lepidoptera, Apidje, and Diptera provided 

 with a proboscis of at least 6 mm. long, and very minute 

 insects of the genera Thrips and MeHgethes, will visit the 

 two forms of / 'iola tricolor for honey, and that, besides, 

 some other insects provided with shorter probosces will 

 seek for their pollen. By direct observation this supposi- 

 tion has been thoroughly confirmed, as shown by the fol- 

 lowing list of visitors actually observed : — 



I. As visitors of the large-flowered form, there 

 have been observed : — {a) Lepidoptera : (i) Picris rapo" 

 L.* (i2).f — {h) Apida; : (2) Boinbus inuscorum L.* (10-15) \ 

 (3) B. lapidarius L.?J (12-14); (4) j5. .y^.§ ; (5) Antho- 

 phora pilipis F.jj (19-21); (6) Andrena albicans K. (J 

 (2-2i),in vain seeking for honey.J — (6-) Diptera : (7) Rkin- 

 gia'rostrata L.§ (11-12) ; (8) Syritta pipieiis L. (2-3), 

 eating poUen.J— (</) Thysanoptera : (9) ThripF.^I 



II. As visitors of the small-flowered form, there have 



» By W. E. Hart (Natuee, vol. viii. p. isi). 



t The numbers enclosed behveen parentheses after the names of the in- 

 sects indicate the length of their probosces in millimetres. 



t By mysell ("Befruchtung der Blumen durch Insecten," p. 145). 



§ By Ch. Darwin, who writes me, May 30, 1S73 :— " Between twenty and 

 thirty years ago I observed, for two or three years, large beds (of V. trico- 

 lor) in the flower-garden, and saw several times Rhingin rostrata, and a 

 nearly black humble-bee visit and fertilise the flowers. I say fertilise, be- 

 cause I had watched the flowers for a long time previously, and saw no m- 

 sect visit them ; but two or three days after the above visits a multitude of 

 flowers withered and set capsules." 



II By Delpino ("Ulterior! osservazioni," p. 62). 



H By Sprengel (" D.is cntdeckte Geheimniss," p. 397), and Mr. A.W. Ben- 

 nett (Nati'KE, vol. viii. p 40). 



been observed : — {a) Lepidoptera : (i) Pieris rapes L. ;* 

 (12), repeatedly; (2) P. vapi L.* (11), repeatedly; (3) 

 Polyotnmatus Dorilis Hfn. * — {b) Apidas ; (4) Apis mclli- 

 fica L. (6) $ ;t (5) Bonibiis hortoriim L.?* (18-21), per- 

 severingly visiting the flowers for honey, although every 

 flowerisdrawndownbytheweightofthis large humble-bee; 



(6) B. Rajclliis FlI.?* (10-13), the same individual visiting 

 sometimes V. tricolor, sometimes Laiiiiuin piirpiireum ; 



(7) B. ?nuscoruin L. (agrorii)u F.) $ (10-14), visiting, with- 

 out distinction, now the flowers of V. tricolor, now the 

 nearly equally large and equally coloured flowers of 

 Litliospcrmum arvense, while omitting the smaller ones 

 of Capsella busra-pastoris, Valerianclla olitoria, and 

 Myosotis versicolor ; (8) Osmia rufa L. (J * (7-9), but 

 once hastily visiting a flower for honey. — ic) Diptera ; (9) 

 Rhingia rostrata L.* (i 1-12), several specimens, repeatedly- 

 visiting flowers for honey. — {d) Coleoptera ; (10) MeH- 

 gethes''' crawling into the flowers. 



Direct observation has thus shown that no essential 

 difference e.\ists between the fertilisers of the large and 

 those of the small-flowered form. But it must appear 

 a striking fact that not only an equal number of different 

 species, but even one more species has been observed on 

 the small than on the large-flowered form. All the visitors 

 of the small- flowered fonn, with the exception of only one, 

 having been observed by myself, I must add, as an ex- 

 planation of this fact, that I have repeatedly watched 

 at the most favourable weather, for several hours, a 

 neglected field, in which, besides some other weeds, 

 there grew an abundance of vigorous specimens of the 

 small-flowered form of Viola tricolor; whereas I have 

 never had an opportunity of watching the large-flowered 

 form under favourable conditions. Therefore I have no 

 doubt that, in spite of the incomplete observations 

 hitherto made on this subject, the more conspicuous 

 flowers are in this species also really far more fre- 

 quently visited by insects than the less conspicuous ones. 

 Otherwise the diflerences in structure and development 

 of the two forms now to be described would be quite in- 

 explicable. These differences are : — i. In the large- 

 flowered form the stigmatic cavity lies somewhat more 

 towards the top of the skull-like end of the style than in 

 the small-flowered one (as shown by the comparison of 

 Fig. 17 with Fig. 18, anil of Fig. 19 with Fig. 20.) 



(i) When the skull-like knob in the two forms is 

 pressed against the lower petal, in the large-flowered form 

 the opening of the stigmatic cavity is directed outwards, 

 so that pollen-grains which have fallen out of the anther- 

 cone spontaneously can never fall into the stigmatic 

 cavity unless carried by insects ; whereas in the small- 

 flowered form the opening of the stigmatic cavity is 

 directed inwards, so that pollen-grains falling out of the 

 anther-cone spontaneously, fall directly into the stigmatic 

 cavity. 



(2) In the large-flowered form the opening of the stig- 

 matic cavity {st, Figs. 17, 19, 21) bears on its lower side, 

 as discovered by Hildebrand, a labiate appendage 

 (/, Figs. 17, 19, 21) provided with stigmatic papillae, so 

 that a proboscis inserted into the flower, when charged 

 with pollen of a previously visited flower, rubs off this 

 pollen on to the stigmatic lip (/), thus regularly effecting 

 cross-fertilisation ; whereas, when withdrawn out of the 

 flower, charged with its pollen, the proboscis presses the 

 lip (/) against the stigmatic opening (st), thus preventing 

 self-fertilisation. This nice adaptation to those visitors 

 provided with a long proboscis (Lepidoptera, Apidse, 

 Rhingia) is completely \vanting in the small-flowered 

 form (Figs. 18, 20, 22). 



(3) In the large- flowered form there is a black wedge- 

 shaped streak {tv. Figs. 17, 19) on the front side of the 

 style, to which Mr. A. W. Bennett first called atten- 



» By myself (June 1873)- 



t By Sprengel (Ice. cit.) and by myself (Ju 

 isiling the flowers for honey. 



75), persi 



ingly 



