RELATIVE PREVALENCE OF AUTOGAMY. 399 



it is probable that more thorough investigation will result in the addition of many- 

 more instances, especially amongst tropical plants. In most cases the species of one 

 genus produce only two forms of flowers; but there are also genera — such as Linum 

 and Oxalis— in which some of the species develop long-, mid-, and short-styled 

 flowers, others long- and short-styled forms, and others again none but flowers 

 with styles of equal length. The determination of the point as to whether hetero- 

 stylism exists or not in a particular case is, in many species, attended with some 

 difficulty, owing to the stamen-filaments increasing in length during the period of 

 the flower's bloom in both the long-styled and the short-styled flowers — a circum- 

 stance which greatly complicates the relations subsisting between the two forms in 

 respect of the lengths of their different parts. There is also some danger of mis- 

 taking for heterostyled species a class of forms which do not in reality come under 

 that category. In the species alluded to, a proportion of the individual plants 

 produce apparently hermaphrodite flowers, with ovaries, styles, and stigmas which 

 can be clearly identified as such, but which nevertheless are not capable of under- 

 going fertilization. 



The results of the investigations into the subject of autogamy recorded in this 

 chapter may be summed up as follows. In plants whose flowers are hermaphrodite, 

 but neither cleistogamous nor heterostyled, both cross- and self-fertilization occur in 

 one and the same flower at different epochs; in plants with cleistogamous flowers a 

 division of labour is established between two kinds of hermaphrodite flowers, of 

 which the one form opens and is adapted to heterogamy, whilst the other remains 

 closed and can only result in autogamy; and, lastly, in heterostyled plants, each 

 species includes two or three different forms of individual, varying in respect of the 

 structure of the flowers, which in the one case aim at cross-fertilization, and in 

 another especially at autogamy. 



In view of the detailed consideration which the methods for promoting autogamy 

 in various plants has received in the foregoing pages, it may not be without interest 

 to allude here briefly to the relative prevalence of this mode of pollination in certain 

 Floras. During the passage of the present edition of this work through the press, 

 a notable addition to our knowledge of Floral Biology has been made b}^ E. Loew 

 (Bliltenbiologische Floristik), in the form of a treatise wherein are summarized the 

 vast number of observations upon flowers and their relations to insects, &:c., so far 

 as the Floras of Europe and Greenland are concerned, that have been published in 

 one place and another during the last ten years. This tabulating of observations 

 has enabled the author to make many interesting comparisons between the Floras 

 of various regions, and, supported by statistics, to exhibit the relative prevalence of 

 types adapted to this or that method of pollination. Though many of the results 

 do but confirm views already the common property of Biologists, they have an 

 altogether special value from the manner in which they have been obtained. 



As regards autogamy, it appears from statistics that it shows an increase in high 

 Alpine forms as compared with plants from a lower level. The accompanying 



