staminodes at whorl IV, and tepals whole or partially persistent on the rim of the cupule, is 
diagnostic. Most of the species can be recognized by this set of characters as belonging to 
Cinnamomum. However, the species that present pinninerved leaves without domatia, are closer 
morphologically to either Persea or Ocotea. On the one hand, species like C. amplexicaule, C. 
chavarrianum, C. erythropus, C. quadrangulum, C. subsessile, C. uninervium, and C. velvetii, 
differ from Persea only in that they have a distinct hypanthium. On the other hand, C. 
formicarium differs from Ocotea only by the presence of well developed staminodia. Even C. 
hammelianum and C. napoense could be considered Ocotea species if not for the presence of 
well developed staminodia and tepals partially persistent in the cupule. Since we don't have any 
indication of how the character syndrome observed in Cinnamomum was attained during 
evolution, we can not be sure that species with those traits conform a single defined clade; the 
same remark is also true for the genera Persea or Ocotea mentioned above. Therefore, in one 
case we can see the Cinnamomum species referred before as Persea species that have grown 
hypanthium or, in the other case, as Ocotea species that have well developed staminodes. In fact, 
there is a group of species recognized in Ocotea that have perfect flowers and glandular 
staminodia (O. indecora group; Rohwer, 1986), which stress a possible connection of these two 
genera. 
In addition to the characters listed above for Cinnamomum, the presence of domatia is a 
feature also found in several Ocotea species, while stamens of whorl III with two sporangia 
occurs in some Persea species. Although in these instances it seems to be a matter of parallel 
evolution, for the core of diagnostic characters of these genera (Cinnamomum, Ocotea, and 
Persea) persists in those species that diverge in the presence of domatia or lower number of 
201 
