242 JOURNAL OF THE ARXOLD ARBORETUM [vol. i 



tion of cunii)lior was 0,877,297 lbs. and 6,935,285 lbs. respcotively. In the 

 fiscal year 1917 (the last return I have seen) the production of camphor 

 is given as 7,371,574 lbs. and camphor oil as 0,727,512 lbs. These figures 

 show that tlic industry has remained about stationary but the tendency 

 now is toward a decline which must increase until plantation camphor 

 becomes available. The quantity of camphor sold in the fiscal year 1917 

 is given as 9,308,900 lbs. valued at $3,400,150 gold; of camphor oil as 

 2,531,819 lbs. valued at $244,901 gold. 



TIIK A:\n:RICAN and ASIATIC SPECIES OF SASSAFRAS 



Alfred Rehder 



The genus Sassafras stood for a long time as an example of a very distinct 

 monotyplc genus peculiar to the flora of eastern North America, until com- 

 paratively recently, in 1907, a Chinese species, Sassafras tzumu, was added 

 to it by Ilcmsley, This Chinese species, though it is so similar in its general 

 apj)earance, in its inflorescence and fruit that without close examination it 

 is difficult to distinguish from the American S. officinale, difTers in its floral 

 structure in several hni)ortant particulars, and for this reason was made by 

 H. Lecomte the type of a distinct genus under the name Pseudosassafras. 

 The chief differences are tlie hermaphrodite or ai)parently hermaphrodite 

 flowers, the presence of a fourth staminal whorl consisting of three stumi- 

 nodes and the ])ubescence on the inside of the base of the perianth. Now a 

 third closely related species appeared when Mr. Wilson, while arranging his 

 extensive collection of the plants he brought back from Formosa, drew my 

 attention to a plant described by Hayata as Lindcra randaicnsis, which 

 looked almost exactly like the Chinese Sassafras tzumu. On closer examina- 

 tion this similarity extended even to the more minute structure of the flower 

 except that the anthers had only 2 locules as correctly descril)ed by TIayata 

 for his Lindcra randaiensis. Instead of 4, as a true Sassafras should have. 

 The number of locules lias been considered by most botanists who have 

 dealt with Lauraceae a very important character, particularly by Pax who 

 bases the main division of the whole family on this character; Benthain & 

 Hooker, Ix^comte and others lay much stress on this character, which, how- 

 ever, seems to lead to an artificial classification. Si)ecies with ^-loculed 

 anthers may occur in genera with normally 4-loculcd anthers, as in Pcrsca 

 cuncata ]\lcissner; and in Persca § Heterandra the anthers of Ihe third 

 series of stamens have two locules, instead of the normal four of the first 

 two series; the same is the case in Phoebe § Ileteranthera. Also the reduc- 

 tion of hermaphrodite flowers to dioecious flowers does not necessitate 

 generic separation, as we have both kinds of flowers in genera like Ocotea, 

 Aydeiulron, Cinnamomum and others, nor does the absence or abortion of 

 a staminodial wliorl necessitate generic separation, as the presence or ab- 

 sence of stamlnodes varies in many genera. Instead of ushig the structure 

 of the anthers whether introrse or extrorse, 4-celled or S-celled, as the chief 



