508 ON THE ORGANS AND:\I01)E OP FECUNDATION
of the stigma, and as a great proportion of the mass so applied is acted upon by the fluid in which it is immersed, the tubes produced are generally very numerous, and to- gether form a cord which passes through the channel of the stigma or style.
On reaching the cavity of the ovarium this cord regu- larly divides into three parts, the divisions being closely applied to those short upper portions of the axes of the valves which are not placentiferous; and at the point where the placenta commences each cord again divides into two branches. These six cords descend along the conducting surfaces already described when speaking of the unim- pregnated ovarium, and generally extend as far as the placentae themselves, with which they are thus placed nearly but perhaps not absolutely in contact.
The cords now described, both general and partial, seem to me to 1)0 entirely composed of pollen tubes, certainly without any mixture of the utriculi of the stigma, or, as far as I can ascertain, of the tissue of the conducting sur- faces.
In two cases, namely Ophrys ajnfera and Cypri2')edium, spectabile, I at one time believed I had seen tubes going off laterally from the partial cords towards the placentae and mixing with the ovula; but I am not at present entirely satisfied with the exactness of these observations, and I have never been able to detect similar ramifications in any other case.-^
That the existence of these tubes in the cavity of the ovarium is essential to fecundation in Orchideae, can hardly be questioned. But the manner in which they operate on, 707] or whether they come actually in contact with, the ovula, are points which still remain undetermined.
I am aware that Professor Amici/ who discovered in several plants the remarkable fact of the penetration of the pollen tubes into the cavity of the ovarium, and Avho re- gards this economy as being very general, likewise believes that in all cases a pollen tube comes in contact Avith an
^ See Additional Observations. = Atnial. des Sc. Nat. xxi, p. 329.
�� �
