BROMELIADS — FOSTER 361 



flowers are present it is almost impossible to identify them correctly. 

 Being semisucculent and very efficient xerophytes, these plants with- 

 stand almost incredible conditions. In certain sections these formid- 

 able terrestrial bromeliads grow in such profusion that it is almost 

 impossible to climb the rocky slopes, for the plants are as well armed 

 with spines as any cactus I know — in fact, they are often mistakenly 

 called cacti. 



Many of these extreme drought-resisting species, like most of the 

 cacti and other succulents, have endured adverse conditions for so 

 many centuries that such conditions have become normal for them — 

 adverse only from our point of view. They are conditions under 

 which they thrive, and should the plant fall from a ledge or a tree 

 to a moist, cool, shady spot, it would probably die. If it did not, its 

 growth would be weak and abnormally fast. They have developed 

 hardy qualities and are seldom found in the soft, shady places where 

 the more tender ones such as Vriesias, Nidulariums, Neoregelias, or 

 Billbergias seek cloister. 



Dr. Smith believes that Puya is the most primitive bromeliad, and 

 he is convincing in his argument against Mez's contention that the 

 most primitive bromeliad is Navia. He suggests that probably Puyas 

 came into being in the high Andes and that their offspring, meeting 

 new situations, produced the various other genera. I, too, surmise 

 that Puyas originated in the territory that is now the Andes, but I 

 suggest that they came into being before the Andes rose to their pres- 

 ent height, and that as the environment is presumably responsible for 

 creating the various genera, they developed from ancestors that have 

 since become extinct. 



But what about the genera, morphologically very close to Pmja^ that 

 are now on the eastern edge of South America, such as Gottendorfia^ 

 L'nchoUrium, or PrionojjhyUum now isolated on the Atlantic coast of 

 Brazil? Finding primitive types of bromeliads so far from their 

 "parents," the Puyas, seems to indicate that in early ages many of 

 these genera perhaps did not evolve from the Puyas but developed 

 simultaneously as a result of their environment. EnchoUnum, Lirid- 

 mania^ Deuterocohnia, and Cottendorfia are similar in construction to 

 Puya^ but that does not necessarily mean that they descended from 

 Puya; they could have evolved from other ancestors now extinct. 



During the period of our two extensive trips into Brazil we collected 

 in three extremes of country which produced the terrestrial species 

 morphologically nearest to the primitive species of Puya. In the 

 Matto Grosso on the Bolivian border we were as close to the "source" 

 as our trip permitted. 



Rising out of the vast marshes in southern Matto Grosso was the 

 strange mountain Urucum, 2,000 feet high and 75 percent manganese 

 ore. Dry areas were always presenting themselves in unexpected 



