As at 2006 there are 4 varieties
var. latifolia
var. major
var. divaricata
var. leucophylla
It is very difficult to do a key now that T. gayii is treated as a synonym of T. latifolia var. divaricata because the rhachis is said to be slightly visible for T. gayii but not for T. latifolia. Clearly var. major with floral bracts to 23mm long should be easily identified because of its large size. There also the comment made by Gilmartin in 1972 that many plants have spreading spikes at anthesis and yet had suberect erect spikes prior to anthesis. One wonders when the various specimens quoted were collected! In cultivation the dazzling white leaves of var. leucophylla are not as prominent making it difficult to differentiate from var. divaricata.
Detail from Gilmartin 1972
The degree of spreading of the spikes in this variety of divaricata, T. latifolia is partially dependent upon the time of collection. When in bud, the spikes are all erect, see illustration. At anthesis or thereafter, the spikes spread at nearly 90° to the main axis of the inflorescence. 'The indumentum is variable and from observations of the specimens collected to date, there seems to he a continuum from the coastal forms with densely gray, lepidote inflorescences, to the Andean forms with moderately lepidote or nearly glabrous inflorescences. The strongly lepidote form extends well up the Rio Chanchan valley and is also near Portovelo in the Rio Grande valley. nExtremely variable in this species is the relative length of the scape compared to the leaf-rosette. AJG 651, a mature specimen from ca. 14 km south of Manglaralto on the coast has its scape exposed about 20 cm beyond the leaf-rosette. Also from the coastal area, near Colonche is AJG 856, a mature specimen with its short scape nearly totally obscured by the leaf-rosette. The same variability of scape-length occurs in the Andean representatives of the variety; for example, AJG 1155 from the Province of Azuay, with a very long scape and AJG 1105 from the Province of Chimborazo with its scape mch exceeded by the leaf-rosette.
In specimens from both the coast and the Andes the holdfasts are well developed and usually have round or elliptic nodules which are probably insect produced. They are similar to the nodules found on the holdfasts of T. disticha H. B. K. and are mentioned in the notes under that species. —SeeSmith & Downs 1977
