yunc 17, 1880] 



NATURE 



149 



this point in his description in the Bot. Ma^., and it was 

 for the further investigation of the case that I raised a 

 plant in my garden. 



The flowers grow on long peduncles, which generally 

 have a horizontal position, projecting some five or six 

 inches from the mass of the foliage. When the calyx 

 opens, the filaments as well as the style are irregularly 

 twisted; but in about two or three days all become 

 straight. The style hangs obliquely downwards ; the 

 filaments all bend sideways, the bend being inside the 

 tube of the corolla, a little over the hairs at their base. 

 There is often a distance of 15 centimetres between the 

 anthers of either side. About 5 or 6 o'clock p.m. the 

 anthers burst, and soon after the style rises and assumes 

 a central position, so that there is a distance of about 10 

 centimetres between the stigmata and any of the anthers. 

 0/i/y then is nectar being secreted by the glandular disk 

 round the base of the ovary, but so copiously that by 

 means of a small pipette I obtained from each flower a 

 mean quantity of o'i4 cubic centimetres. This nectar is 

 completely transparent, very sweet, and slightly mucila- 

 ginous. It contained a kind of gum which is precipitated 

 by absolute alcohol. Tlie nectar appears therefore -when 

 the anthers have done their -worhj even an hour before 

 their rupture no trace of it is to be found. The nectar- 

 cavity in the tube of the corolla is completely shut up by 

 the numerous spreading hairs at the base of the filaments, 

 so that an outflow is impossible. The grains of pollen 

 are very large (o'2 millim. in diameter) and of the same 

 structure as in Cobeea scandens. They are covered by a 

 glutinous layer, and are heavier than water. 



Several weeks passed at first before I witnessed the 

 manner of fertilisation. The stigmata were every morning 

 ■carefully examined, but no pollen could be discovered on 

 them. The filaments twisted back again and got some- 

 what frizzled, after one single night's e.xpansion. About 

 noon the corolla drops off, separating from close to the 

 glandular ring, and then slipping down over the style, 

 which, by this time, is again in a relaxed hanging position. 

 There is always some nectar in the tube of the corolla 

 after its separation, but none remains in the calyx round 

 the ovary, nor does its secretion continue. 



These facts show clearly that the fertilisation must take 

 place in the same night after the bursting of the anthers, 

 and it was but natural to suppose that it was effected by 

 nocturnal moths. It would appear, furthermore, that the 

 nectar /; not of any direct advantage to the plant, as Mr. 

 G. Bonnier emphatically affirms {Annales des Sci. A^at. 

 Bot., ser. vi. vol. viii. p. 206), because of its being pro- 

 duced and lost in all flowers, fertilised or not, in the same 

 ■way. 



As soon as the number of flowers increased (on some 

 evenings twenty to twenty-five had their anthers opened), 

 I found every morning most of them with pollen on the 

 stigmata, and keeping a close watch, I discovered that 

 the plant was visited by several large Sphingida: belong- 

 ing to the genera Chaerocanipa, Diliidia, and Amphonyx. 

 I observed altogether four visits of an Amphonyx, three 

 of a Chaerocampa, and one of a Diludia. All of them 

 proceeded in the same manner. Holding the body close 

 over the style, they dipped their spiral tongues into the 

 tube of the corolla, beating all the while the anthers so 

 violently with the tips of the fore-wings that they dangled 

 about with great velocity in every direction. The grains of 

 pollen being covered by a sticky substance, many of 

 them adhered to the wings. I have caught an Amphonyx 

 which, after having visited six flowers consecutively, had 

 the tips of the fore-wings almost yellow with pollen. 

 When leaving a flower for another one, some of this 

 pollen is even lost on the foliage, but by the time the insect 

 takes its central position before the flower the stigmata 

 are likewise touched by the wings, and thus some pollen 

 is left on them. Some flowers remain without being fer- 

 tilised, especially in places where the moths cannot reach 



them easily. All flowers fertilised in this manner set 

 fruit very soon ; but no flower gave a fruit without having 

 its stigmata poUenised by crossing. 



Self-fertilisation is therefore excluded, and this is 

 further proved by the following experiments : — Twelve 

 flowers were artificially fertilised by their own pollen and 

 afterwards protected by muslin bags ; only in one case 

 was a fruit obtained ; but I am not quite sure whether 

 there did not come some foreign pollen on the stigmata of 

 this flower. Cross-fertilisation was likewise tried in twelve 

 flowers, nine being experimented on in the same even- 

 ing after the opening of the anthers, and three the next 

 morning. All the former are now with fruit ; the latter 

 remained sterile. This fact shows how very short is the 

 period of possible fertilisation. 



Flowers visited by nocturnal moths are as a rule either 

 large and of white colour, or have a strong smell ; but in 

 our Cobcra the former is certainly not the case, and my 

 olfactory nerves at least cannot discover any smell. But 

 it is well known that insects, especially Lepidoptera, are in 

 this respect of a really wonderful keenness, which enables 

 them to track a scent absolutely imperceptible to man. 



As I shall have a considerable crop of Cobasa-seeds, I 

 can ofter some to any botanists who should wish to grow 

 the plant. A. Ernst 



Caracas, April 4 



P.S. — As soon as^the corolla has fallen off, the peduncle 

 withdraws slowly amongst the dense foliage, where the 

 fruit develops, protected from all kinds of injury. 



EXPERIMENTAL RESEARCHES IN 

 ELECTRICITY^ 

 Part III. — Tibc-Potentialj Potential at a Constant 

 Distance and Various Pressures ; Nature aud Phe- 

 nomena of the Electric Arc. 

 T\/rESSRS. De La Rue and MiiUer, in the third part of 

 •'■'•^ their researches on the electric discharge, commence 

 by describing a series of experiments to determine the 

 potential necessary to produce a discharge in a tube, ex- 

 hausted gradually more and more while using a constant 

 number of cells in all the experiments. In consequence of 

 the life of the battery becoming so much exhausted by the 

 method employed the experiments were confined to one 

 gaseous medium, namely, hydrogen. Since the completion, 

 however, of the measurements described in the paper the 

 authors have found two other more convenient methods 

 for determining the lube-potential, which do not exhaust 

 the battery injuriously ; these are described in an appendix. 

 The tube, 162, employed was 33 inches long and 2 inches 

 in diameter, the distance between the ring and straight 

 wire terminals being 2975 inches; the battery consisted 

 of 11,000 cells. The discharge took place when the 

 pressure was reduced to 35'5mm., 46,710 M (millionths 

 of an atmosphere), and the exhaustion was afterwards 

 continued gradually until it fell to o'oo65 mm., 8'6 M- 

 In commencing each set of experiments the deflection of 

 a tangent-galvanometer was observed when the battery 

 was short-circuited. By a table previously calculated the 

 value of the deflection in ohms of resistance per cell could 

 be read oft"; this, multiplied by 11,000, gave the total 

 resistance of the batter)- ; the tube was then connected 

 with the terminals and the galvanometer again observed ; 

 this gave a less deflection and indicated a greater resist- 

 ance, which, multiplied by 11,000, gave the total resist- 

 ance of the tube and battery : by subtracting the resistance 

 of the battery the resistance of the tube was ascertained. 

 Calling the total resistance R, the tube resistance ;-, the 



tube-potential V, V ^ 



X 11,000 



IT 



The tube-potential re- 



quisite to produce a discharge, with a pressure of 46,710 

 M, was found to be 10,250 cells ; thi-i gradually fell until 



I ■' Experiment.1I Researches on the Electric Discharq:e with the Chloride 

 of Silver Battery." by Warren De La Rue. M.A.. D.C.L., F.R S., and 

 Hugo W. MuUer, Ph.D., Y .K.S.i.P/iil. Trans., vol. clxxi. p. 65). 



