IQO 



NATURE 



\July 8, 1875 



column of water would indicate directly the intensity of 

 the radiation of the sun in calories if the ice did not also 

 partly melt in consequence of the surrounding warm air. 

 In order to eliminate this influence, the progress of the 

 column of water must be observed before and after the 

 actual experiment, and during these observations the sun's 

 rays must be shut out from the apparatus by a screen. 

 The difference of the readings with and without the sun's 

 rays will then indicate the density of the latter. But this 

 method has a drawback. It was found that with experi- 

 ments which were made in quick succession, when the ap- 

 paratus was exposed to the sun's rays, that the first results 

 were always a little larger than the following ones, and 

 that only after some time had elapsed did the results show 

 a constant value. The reason of this is doubtless the 

 formation of a stagnant layer of water in the apparatus 

 below the blackened plate, and this layer must first reach 

 a stationary position before anything like regularity is ob- 

 tained in the results. 



With regard to the general results of these experiments; 

 which were made by Messrs. Rontgen and Exner on the 

 platform of Strassburg Cathedral, the absolute values of 

 the intensity of the radiation of the sun are considerably 

 larger than those found by Pouillet, If Pouillet's values 

 are reduced to the same measures and units, which form 

 the basis of the values obtained by Rontgen and Exner, we 

 find, for instance, for the month of June and the sun's eleva- 

 tion 12I1, the value i'i40, while the latter observers still 

 obtained i"226 for an elevation of I2h, 15m, Further, we 

 must remark that the values obtained by Rontgen and 

 Exner are decidedly too small (the observations record 

 the progress of the column of water after the stationary 

 condition of the stagnant layer of water), and that accor- 

 ding to a rough guess they should be at least 20 per cent. 

 to 25 per cent, larger ; thus it is certain that Pouillet's 

 values must be looked upon as considerably too small. 



FERTILISA TION OF FLO WERS B Y INSECTS* 



XI. 

 Adaptation of Floivers to Lepidoptera — Hesperis tristis. 



LEPIDOPTERA are distinguished among all insects 

 that visit flowers by their slender proboscis. Hence, in 

 order to make their honey exclusively accessible to these 

 insects, flowers have only to narrow the entrance to their 

 nectaries to such a degree that no other proboscis but 

 that of Lepidopterous insects is able to enter. This 

 adaptation to butterflies by narrowing the entrance of 

 the nectary in different families of plants has been 



Fig. 65.— Flower oi Hesperis tristis (natural size). 



arrived at in very different ways. In flowers with a 

 tubular corolla {Primula villosa, Daphne striata. Nature, 

 vol. xi. p. no. Figs. 43-47) the corolla- tube has narrowed; 

 in flowers with a honey-secreting spur {Gymnadenia, 

 Ni^ritella, Nature, vol. xi. p. 170, Figs. 58-62) the 

 entrance of the spur has been constricted ; in the labiate 

 -flowers of Rhinanthus alpitms (Nature, vol, xi, p. in, 

 Figs. 51-56) the large entrance of the flower is blocked up 



* Continued from p. 50. 



by the margins of the upper lip lying close together, and 

 only a small opening in its rostrate projection has been 

 left open ; in the quite open flowers oi Lilium Martas;on 

 (Nature, vol, xii, p, 50) the honey-secreting furrow at 

 the base of the sepals and petals has been converted into 



Fig. (ye. 



Fig. 66.— The same after the sepals, the petals, and two of the four longer 

 anthers have been removed. «, nectary ; h, honey ; a', shorter anther ; 

 st, stigma. 



Fig. 67.— Situation of the nectary, aa, longer filaments ; o, point of inser- 

 tion of one of the shorter filaments ; b b, points of insertion of the two 

 adjacent petals ; d, insertion of the adjacent sepal ; n, nectary. 



a narrow channel by a coating of glandular hairs, Hesperis 

 tristis, belonging to the family of Cruciferae, which are 

 generally visited for honey by Apida;, Syrphidse, Muscidse, 

 and various other insects, has excluded from its honey all 

 visitors except Lepidoptera, by simply lengthening its 

 sepals and the basal portion of its petals and laying them 

 close together. The sepals, indeed, as is shown by Fig. 



.jA\\\m 



Fig. 68. — The centre of the flower at its first period seen from above. 

 a, longer anthers ; o, openings. 



65, are elongated to 11-15 rn^i-, and whilst diverging 

 and presenting open slits in their basal portion, are con- 

 vergent and connate towards their tips. By this coal- 

 escence of the sepals the entrance of the flower is so 

 constricted as to be almost completely filled up by the 

 four longer anthers (a, Figs. 68, 69). At first, when the 



