6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I26 



pHshed by giving the unadorned name to the locahty with the best 

 title to it, and modifying the name or completely changing it for the 

 others. The words "do Norte" are added to the northernmost locality 

 of that name and so on for the points of the compass, others are 

 compounded with the name of the river or mountains which they 

 adjoin. On the other hand, one state capital goes from Paraiba to 

 Joao Pessoa without a backward glance. Ultimately this latest reform 

 should clarify Brazilian geography greatly, but at the moment it is 

 an additional confusion to one dealing for the most part with the 

 older names for localities. 



The recent practice of dividing the states into municipios helps with 

 the identification of many localities, and the Tabuas Itinerarias 

 Brasileiras (Servigo Grafico do Institute Brasileiro de Geografia e 

 Estatistica-1950) is a very useful index to the municipios. I have 

 also been fortunate in obtaining the latest maps for a few states, 

 notably Parana and Santa Catarina. 



When all other resources have been used, I have appealed to the 

 collectors and they have helped me with their own localities and often 

 with those of past botanists. Dr. E. Asplund of the Riksmuseet, 

 Stockholm, has been most helpful in finding the localities of Swedish 

 botanists in Brazil. In order to save anyone the work of identifying 

 these localities again, on page 259 I have listed those which are not 

 explained in the "Index to Map of Hispanic America i : 1,000,000." 



In citing localities in the text, the sequence is from the vague to 

 the definite. First come the collections of which we know only that 

 their origin was Brazilian, next come those that are located only as 

 to state, followed by those with named localities within the state, 

 and, finally, by those that are further defined by a municipio. Locality 

 names which are defined only as far as the state are cited in the text 

 in parentheses, indicating that it has not been possible to plot them 

 on the map. 



DISTRIBUTION AND ORIGIN 



The map on page 7 (from Goode's series) shows the approximate 

 position of all bromeliad localities in Brazil that it has been possible 

 to identify. The observation that range maps represent where collec- 

 tors have been rather than where plants are, is peculiarly apt in this 

 case because of the difficulty of collecting and preserving bromeliads. 

 Undoubtedly this situation accounts for the marked breaks in repre- 

 sentation in eastern Brazil around the Baia-Espirito Santo boundary 

 and around that between Sao Paulo and Parana. 



