32 DAVENPORT ACADKMY OK NATURAL SCIENCES. 



The attempt made at an early day by Nuttall, and later by Klotzsch, 

 to se\)arate Arctostaphylos into several distinct genera, founded on dif- 

 ferent fruit characters, though not generally adopted, has been very 

 properly used in systematic botanical works to divide the genus into 

 natural sections. 



Much of the difficulty in properly discriminating species — especially 

 such as, from their extensive geographical range, are subject to extreme 

 variations in thei/ exposure to different conditions of soil and climate — 

 is due to the fact that ordinary collections do not include identical 

 specimens in different stages of flower and fruit, which, as in the pres- 

 ent genus, are often separated by a considerable interval of time. The 

 only proper remedy for this will be found in continuous field observa- 

 tions covering the whole period of growth. In the following paper I 

 have endeavored to bring together some of the results of such obser- 

 vations, as far as my recent opportunities have afforded, to complete 

 or rectify our knowledge of Pacific coast species. 



The great importance of fruit characters to properly distinguish, not 

 only genera, but species, becomes especially apparent in making a 

 careful examination of any of our imperfectly known plants. Thus, 

 in the genus under consideration, the uniformity of floral characters in 

 everything but size, color, and degree of pubescence, affords absolutely 

 nothing on which to base specific characters ; what applies to one will, 

 in great measure, ap])ly to all, within the limit of ordinary variation. 

 The leaves, stems, and inflorescence afford better-marked characters, 

 but cannot be relied on in distinguishing such variable si)ecies as A. 

 pungens, A. ioi7ie?itosa, and the doubtful ones that have been variously 

 referred to one or the other of these. 



In a somewhat extended examination of living plants, and frequent 

 dissection of all the accessible fruit, connected, as far as possible, with 

 flowering specimens from the same bush, I find little difficulty in dis- 

 criminating species by the fruit alone, except in the following sec. 

 tion. 



Section 2. Uva-Ursa, Gray's Synopt. Fl., Vol. 2, p. 27; Daph- 

 nidostaphylis, Klotzsch. 



Here the irregularity in the degree of coalescence of the nudets and 

 the apparent variability in the number of the cefls seem to offer no 

 distinctive characters to be relied on to separate s]jecies, so that it is 

 only by combining other distinct features that they can be properly 

 discriminated. 



