Maté or Brazilian holly (_Ilex paraguayensis_) belongs to the same genus. Its ash analysis, as made by Señor Arate, is given in the second column. The plant grows wild in Brazil, and is very largely used by the South Americans. It has, according to Peckolt (Pharm. J. Trans. (3) 14, 121-124; Abstract Jour. Chem. Soc., 1884, 479), been planted, and seems to succeed well, in the Cape of Good Hope, Spain, and Portugal. It is stated that six different species of Ilex are used in the preparation of this tea. Peckolt gives, in his analysis of the air-dried leaves, the percentage of caffeine as 0.639. The average percentage of analyses by different authors is about 1.3. I can find mention of only one other Ilex used as a substitute for tea. The analysis of this by Ryland and Brown is quoted in Blythe’s “Composition and Analysis of Foods” (p. 343). It is called the Ilex cassiva, is said to be used as a tea in Virginia, and the percentage of caffeine is given as 0.12. This is probably the same thing as the yopon, the analysis of which is given above, and the “cassiva” may be a misprint for “cassine.” In a more recent paper Professor Venable reports additional analyses, which are interesting. He says: Some years ago an analysis of the leaves of Ilex cassine was given in this journal.[4] In this analysis appeared the interesting fact that these leaves contained a small percentage of caffeine. During the winter of 1885-’86, at the request of some medical friends whose attention was drawn to the analysis, a more thorough examination was undertaken, not only of the leaves, but of the berries. It was thought advisable, at the same time, to examine the leaves and fruit of other representatives of the Ilex family in this State--_Ilex opaca_ and _Ilex dahoon_. This was primarily a search after alkaloids, and not intended as a complete chemical examination. As no alkaloids were found other than the caffeine already mentioned, no account of the work was published, and the results have been hidden away in my note books ever since. Thinking, however, that even negative results are often of some value and that the partial analysis might be of aid to others, I offer this paper for publication in the journal of the society.