IV 
PREFACE. 
the production of the celebrated IMate, or Paraguay Tea, a noted article of 
commerce still extensively consumed in Brazil, Paraguay, the several 
Argentine provinces, Chile, Peru, and Bolivia, in the same Avay that the 
Tea of China is used in Europe, I succeeded in obtaining from the re- 
nowned Bonpland, who resided many years in Paraguay, authentic speci- 
mens of these plants, verified by his own handwriting : these typical 
plants, as well as those employed in Brazil, are here described and figured, 
together with some other genera of the Aquifoliacea allied to them. 
The genus Goupia, the structure of which was not well understood, is 
here figimed, with a complete analysis, and its affinities considered. 
All the South-American forms of Ephedra known to me are here also 
figured and described, and careful analytical details of the structure of 
this genus are shown. A review is also given of the opinions of different 
botanists in regard to the structure and affinities of the Gnetacece, and 
numerous facts are offered to invalidate the views of those who place the 
family among true Gymnosperms. I have endeavoured to show that 
Ephedra cannot be regarded as a Gymnosperm under any legitimate point 
of view, particularly as it has neither naked ovules nor naked seeds, and 
as by its organization it can only be considered very distantly allied to 
CojiifercB and Cycadaceae. I have pointed out that it has a calycine deve- 
lopment assuming the form of an involucre, a regular petaloid perigonium 
(or tubular corolla), and determinate raonadelphous stamens in the male 
flowers ; while the female flowers have a similar calycine development, the 
ultimate pair containing two ovaries, with a depressed style and an open 
stigma, each, when matiured, producing a fruit with a coriaceous pericar[) 
enclosing a single albuminous embryonary seed covered by two distinct 
integuments. This organization has ample claims to be considered truly 
exogenous, exhibiting a far higher order of development than the Conifera', 
Cycadaceee, or even Myricacere and other amental orders, approaching in 
these respects nearer to Uriicacea, with which it agrees in many correla- 
tive features, notably in the presence of cystoliths or spicular fibres in its 
