8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 126 



Distrito Federal represents dozens of localities, nearly a hundred 

 species, and great quantities of individuals, while one spot in Minas 

 indicates a single locality as a rule and that with few species and 

 individuals. 



Great general collectors who have operated in the east and also in 

 the interior give further evidence along this line, as shown by a com- 

 parison of A. Glaziou's collections in the state of Rio de Janeiro and 

 in Goias, or those of F. C. Hoehne in Sao Paulo and Mato Grosso. 

 In each case the collections in the east are more numerous, as might 

 be expected from the greater amount of time spent there, but they 

 are also significantly richer in species both totally and in relation to 

 the remainder of the flora. Such collectors in the Amazon basin as 

 A. Ducke, R. L. Froes, B. A. Krukoff, J. G. Kuhlmann, R. E. 

 Schultes, and R. Spruce afford us almost no direct contrast with the 

 bromeliads of the east, but the family is a very insignificant part of 

 their well rounded general collections. 



Finally, we have the evidence of collectors who were interested in 

 bromeliads almost exclusively. Dr. J. L. Collins, who traveled ex- 

 tensively investigating the origin of the pineapple, has told me how 

 driving westward in Pernambuco he saw bromeliads in profusion 

 and then abruptly could find none. The following from the unpub- 

 lished report by K. F. Baker and J. L. Collins of exploration in 

 1938 and 1939 gives the details : 



On the return trip from Candado we had opportunity to determine more closely 

 the exact point at which the Bromeliaceae stopped in the westward distribution. 

 A short distance east of Salgadinho we suddenly came to the end of the Sertao 

 and passed into the curimataii and with this change immediately passed from 

 a situation of almost no Bromeliads to their supplying one of the dominant 

 elements of the flora. 



The Fosters in their journey to Mato Grosso repeatedly found 

 formations that they had learned to associate with bromeliads in the 

 east completely lacking in these plants. 



Concerning the origin of the Brazilian bromeliads there seems little 

 reason to alter the general conclusions reached in my "Geographical 

 Evidence on the Lines of Evolution in the Bromeliaceae" (Bot. Jahrb. 

 66: 446-468. 1934). Using the generally accepted morphological cri- 

 teria, the Brazilian bromeliads are characterized as derived rather 

 than primitive types. The Bromelioideae, with its consistently in- 

 ferior ovary and indehiscent fruit, is easily the most advanced of the 

 three subfamilies and has its chief center in eastern Brazil with a 

 number of endemic genera. All this would indicate that the family 

 had arisen in western South America and migrated into Brazil. 



