THE ORIGIN OF FLORAL APPENDAGES. 139 



of liairs. Similarly, those of Cecidomyia Ulmarice on Spircea 

 TJlmaria are hairy outside, and papillose within ; while similar 

 ones of a Pliytoptus on the Sycamore are lined with long 

 blunt-ended hairs, and are clothed without by others, long 

 and pointed. In all these cases the galls, as well as the 

 hairs, are the product of irritation set up by the presence of 

 the egg deposited by the insect.* 



As another very common instance of the presence of 

 epidermal papillae and hairs, may be mentioned their occur- 

 rence in the stylar and ovarian cavities. The former, and 

 the placentas especially, may be clothed witli delicate hairs 

 exactly resembling root-hairs. Such may be well seen in 

 the Poplar, Tamus, Richardia JEthiojnca, etc.; and since M. 

 Guignard f has discovered that the mechanical and physio- 

 logical irritation of the pollen-tubes is required to cause 

 their development on the walls of the ovary in Vanilla, 

 between the longitudinal bands of conducting tissue, it is, 

 I think, a by no means improbable theory that the tufts of 

 hairs over the nectaries, " tangles," " wheels," etc., on the 

 filaments or corolla- tubes, have been actually caused by the 

 irritation of insects, since they occur just where such irrita- 

 tions are made. 



One use of certain outgrowths has been regarded as 

 intended to protect the honey from rain. Why, however, 

 some flowers should be so favoured while many others, as 

 of the Umhelliferm, have no protection at all, is not stated. 

 The interpretation I have here offered will, of course, apply 

 to all such growths, whenever they may really keep off rain 

 or " unwelcome guests." 



* Krasan has lately discnsFed the formation of the woolliness of 

 galls, etc., Oesferr. Bot. Zeitschr., xxxvii. (1887), pp. 7, 47, 93, seqq. 



t Sur la Pollinisation et ses Effets chez les Orchid^eSy par M. L. 

 Guignard, Ann. des Sci. Nat., torn, iv., 1886, p. 202. 



