^ PREFACE. 



Urban/ in ' Linnsea/ B. vii. p. 621. Mr. Meehan (' Bull. 

 Torrey Bot. Club/ vol. vi. p. 189) has endeavoured to 

 throw doubts on my observations on the sterility of the 

 forms of L. perenne when fertilised with their own-form 

 pollen, because a plant from Colorado yielded seed when 

 growing by itself ; but as might have been expected, and 

 as is sufficiently clear from the remarks of a well-known 

 reviewer in the ' American Journal of Science/ Mr. 

 Meehan mistook L. Lewisil, which is not heterostyled, 

 for L. perenne. 



In the Boragineae, Lithospermum canescens differs, 

 according to Mr. Erwin F. Smith (' Bot. Gazette/ 

 United States, vol. iv., 1879, p. 168), from the hetero- 

 styled species of the same genus by occasionally present- 

 ing a mid-styled form, which has a short pistil like that 

 of the short-styled, and short stamens like those of the 

 long-styled form. All the forms seem variable, and the 

 whole case requires further investigation. 



Mr. Alex. S. Wilson informs me that on comparing 

 the pollen-grains from a long-styled plant of Erytlircea 

 centaurium with those from some short-styled plants 

 from the island of Arran, they differed in size and 

 shape, as in the case of the undoubtedly heterostyled 

 Menyanthcs trifoliata, a member of the same family of 

 the Gentianeae. I had myself formerly observed that 

 the flowers on different plants differed much in struc- 

 ture, but could not make out that they presented two 

 distinct forms. 



The Rubiacese contain many more heterostyled 

 plants than any other family, and several additional 

 cases can now be added. Mr. C. B. Clarke has been 



kind as to send me sketches made in India of two 



extremely distinct forms of Adenosacme longifolia. 



e remarks "that the peculiarity of the case is not 



the difference in the length of the style and stamens 



