AUTOGAMY BY LENGTHENING OF STAMENS. 335 



in which this is observed to occur:— the Moschatel (Adoxa Moschatellina), most 

 of the species of the Knawel (Scleranthus), Pcederota Bonarota, of wide dis- 

 tribution in the Southern Alps, the curious A'ponogeton distachyon, native to 

 South Africa, and a large number of Cruciferae, Saxifragacese, Willow-herbs, 

 Geraniacese, Convolvulacese, and Caryophyllacese. 



Of the large family of the Cruciferae we may mention Arabis coei^lea, 

 Braya alpina, Gardamine alpina, and Rhizobotrya alpina, all of which are small- 

 flowered species growing in the upland hollows of high mountains, and in addition 

 the annual or biennial species named Lepidium campestre, L. sativum, Sisym- 

 brium Alliaria, S. Thalianum, Thlaspi alliaceum and Thlaspi arvense. In these 

 plants the stigma is sessile on the ovary in the shape of a small round cushion, 

 which becomes visible the moment the imbricate petals of the bud begin to 

 move apart. At this period only cross-pollination can take place, as all the 

 anthers in the flower itself are still closed; but the four long stamens now grow 

 up along the wall of the ovary until the anthers are exactly on the same level as 

 the stigma. The anthers have dehisced by that time, and their pollen cannot but 

 be deposited on the receptive cells at the periphery of the stigmatic cushion. 

 Another observation which has been repeatedly confirmed, is that only one of the 

 four anthers parts with its pollen to the adjacent stigma, while the rest, though 

 quite close to the stigma, are not in immediate contact with it. The pollen of 

 these three anthers is apparently placed there so that it may be carried off by 

 the small flies which visit these cruciferous plants and transferred by them to 

 younger blossoms. 



The Saxifrages (e.g. Saxifraga androsacea) here in question have two linear 

 i or oblong stigmas. After the anthers are raised, the pollen is usually brushed off" 

 them on to the sides of the stigma near its base. But here again it is note- 

 worthy that for the most part only one of the five anthers devotes its pollen to 

 the process of autogamy, whilst the others remain a little below the stigmas, and 

 do not come into contact with them. 



In a number of small-flowered Willow-herbs {Epilobiwm collinum, E. montanum, 



E. parviflorum, &c.) the stigma is composed of four thickish divergent lobes 



arranged in a cross with four angles between them. Upon the first parting of the 



petals, which always happens early in the morning, the anthers may be seen to be 



I underneath the mature cruciform stigma, but in the course of that very day the 



filiform filaments grow to a sufiicient length to place the anthers in the re-entrant 



I angles of the cross. Meanwhile, dehiscence has taken place, and by the evening 



i of the first day autogamy ensues. During the night the petals close up, and 



I the flower droops a little; the next morning the petals open again, and it is then 



I evident that the filaments have grown rather longer, two or three pollen-covered 



! anthers standing above the stigma and partially concealing it. The place occupied 



j by the stigma on the previous day is now filled by an assemblage of anthers laden 



i with pollen, which is brushed ofi" by insects and transported to other flowers. 



Thus we find in these Willow-herbs that on the first morning cross-fertilization 



