48 MISSOURI BOTANICAL GARDEN. 



ouping, the species have 



rly as possible (it 

 ) in the order of 



distinguish 



We 



t a. 



S. alba, fragilis andlucida; 6, phylicifolia and argyrocarpa; 

 c, nigra and amygdaloides ; d, adenophylla, discolor, cor- 

 data, petiolaris, sericea, myrtilloides, and, more distantly, 

 Candida; e 9 humilis and tristis; /, rostrata, glaucophylla, 

 and balsamifera. Each of the remaining species, viz. : S. 



longifolia, herbacea, and Uva 



may 



from 



transmitted 



print by the artotypist. In spite of a reasonable amount of 

 patience and effort it will be noticed there are considerable 

 defects. The inherent difficulties in rendering some of the 

 leaves were too great to be overcome. In selecting leaves 

 for the plates from samples, regard was shown for those 

 which should best exhibit the principal venation character- 

 istics, although others might promise better results as to 

 mechanical execution, or make finer pictures, or even present 

 a more characteristic outline. 



The reader is cautioned against supposing the represen- 

 tations to present complete and perfect pictures of what 

 may be seen by transmitted light in the natural leaves. 

 The very minute details, which all help to make up the 

 impression, are, generally, more or less wanting. Even 

 though at the first, the print was quite complete as to min- 

 utiae, something of its definiteness was lost at every step of 

 the successive manipulations ; so that wherever the natural 

 leaf is available, it is to be regarded as far preferable for 

 study. In reference to the terms employed in the descrip- 

 tions, need was felt for some not ordinarily used in botan- 

 ical descriptions. As definite a meaning as possible was 

 given to each one used, and such meaning assigned to each 



as will be found defined in the accompanying glossary. 

 The terms " loop M and M arch," so much used, were found 



3 



