DAEWm OK PEIMULA. 241'. 



braiie and the lower, which separates them from the ovary, aud the 

 anthers, closely applied, are apparently quite connate and together 

 adnate to the stigma. We have observed in soaked specimens what 

 we have every reason to believe are true pollen granules, with their 

 tubes penetrating the tissue of t]ie stigma. "', 



The contents of the ovary do not appear to differ in the normal 

 and abnormal flowers. In Campanula colorata we have seen flowers 

 intermediate in character between those above described, and normal 

 ones, in which the corolla, instead of being imperforate, opened by 

 teeth in the centre, though, falling short of the calyx- lobes in length, 

 — the style considerably lengthened and the anthers free. In con- 

 nection with the occurrence of dimorphous flowers in Campauulaceae, 

 it may be well to bear in mind that the method of fertilization of the 

 noi'mal flowers was long a puzzle to botanists. For a detailed 

 notice of the vai'ious hypotheses suggested to explain it we must 

 refer to M. A. de Candolle's Monographic des Campanulees (1830) 

 and especially to M.M. Brongniart* and Tulasne'sf Papers in 

 " Amiales des Sciences Naturelles." In these flowers the anthers 

 open and discharge their pollen before the expansion of the corolla. 

 M. Du Petit-Thouars conjectured that the stigmas were fertilized 

 before it opened. He found that the stigmatic lobes were slightly 

 divergent in the bud at a time when the anthers might be supposed 

 to open and that they again close shortly before the corolla expands : 

 after its expansion they are once more divergent. This view was 

 considered to be supported by the case of the allied Goodeniete and 

 Scaevolae in -which the pollen is received into a capsule or indusium 

 terminating the style before the flower opens. AVhen the corolla ex- 

 pands the indusium in these plants is closed. Again, much attention 

 has been directed to what have been termed the ' coUecting-hairs ' 

 with which the style in the Bell-flowers is so frequently clothed. 

 A function has been attributed to them in the fertilisation of the 

 flower, but this, as Brongniart showed was due to imperfect observa- 

 tion. These hairs, which brush ofi' the remaining pollen from the 

 anthers as the style shoots up through them, frequently become 

 invaginated, like the finger of a glove drawn back half way up : the 

 sheathing portion entangles a few of the grains so that they appear 

 actually drawTi into the tissue of the style : hence the mistake. M. 

 Tulasne, whose observations are of the highest authenticity, finds 

 that pollen received upon the stigma produces the tubes which ferti- 

 lize the ovules. How the pollen reaches the stigma must be more 

 fully settled by careful observation. It is highly probable that 

 insects play an important part in its conveyance, as various observers 

 have suggested. There are other plants belonging to difterent 

 Natural Orders to those above noticed, which oflfer like dimorphism. 

 In Caryophyllaceae, Maximowicz,J describes a Stellaria (dis- 



* 2e Ser. xii. 244. f 3e Ser. xii. 71. 



X Primitiae Fl. Amurensis, 57. 



