Lasthenia of the Baeriinae, but they are most common in the Ecliptinae and the Coreopsidinae. Sheareria Moore of China, formerly placed in the Milleriinae, is a member of the tribe Astereae.
Subtribe 1. AMBROSIINAE
The non-synanteric condition in the more advanced genera of the subtribe is unique in the family. Such a relationship was not mentioned in the revision of the strain by Stuessy (1977), where Engelmanza. Many detailed studies of Ambrosiinae deal with specializations associated with wind pollination in pollen grain or inflorescence form (Wodehouse, 1928; Payne Skvarla and Larson, 1965; Payne and Skvarla, 1970).
Subtribe 2. ESPELETIINAE
The relationship of Parthenice and Parthenzum with the Ambrosiinae was recognized by Bentham, but the two genera were technically placed in the adjacent part of the subtribe Melampodiinae in the Bentham and Hooker system (1873). This was mainly based on the presence of fused anthers and achene-palea complexes in the two genera. T he Espeletiinae are restricted to fairly high altitudes in the northern Andes, where they are characteristic of the typical Paramo habitat.
MELAMPODIINAE
T h e Melampodiinae have traditionally included most members of the Heliantheae with suppressed gynoecia in the disc flowers but lacking specialization towards wind pollination. Of the taxa included in the subtribe at various times, many have been excluded in recent studies, Philoglossa DC. T e striations in the pains of the lower trunk show variations to a degree that contrasts with most other groups of the trunk.
Subtribe 4. POLYMNIINAE
In Acanthospermum, striations are present only in the form of irregularly spaced lines that usually extend only part of the length of the achene. We would conclude that the striae of Melampodiinae are in a more pleimorphic state than those of most Heliantheae. These provide an important contrast to the genera of Melampodiinae with which Pobmnia has previously been associated.
Subtribe 5. MILLERIINAE
At the outset of this study, Ckxza was considered a member of the Polymniiae, sharing most of the essential features of the subtribe. Polymnza does not have the same characteristics as the Neurolaeninae seen in C'n,tra, but like Chxza it is perhaps closest to genera with hermaphroditic disk flowers. Guizotia and Hefianthus, for which the amino acid sequences of cytochrome c are known, differ subtribally but are apparently in the same paleaceus series of tribes.
DESMANTHODIINAE
Guizotia in the Milleriinae, however, is based mainly on close approximation of technical characters and is made despite differences in habit, strictly yellow rather than white flower color, and colored resin in corolla canals. The genus is unusual in the Heliantheae because of its restriction to the African region and its uncertain relationship to anything in the Western Hemisphere. Differences of 8 amino acids (Boulter et al., 1972) indicate considerable instability in the molecule in the tribe.
Subtribe 7. CLIBADIINAE
PINILLOSINAE
GUARDIOLINAE
Guardtola has usually been placed in the Melampodiinae because of the suppressed gynoecy of the disc flowers. Initial doubts regarding the placement of Stuessy stemmed from the presence of glands on the anthers of Guardzola and the apparent lack of glands on the anthers of Coreopsidinae (Robinson and Brettell, 1973a). On further analysis, Guardzofa shows distinct striations in the achene wall and heavily ornamented cells in the seed coat, characters also lacking in Coreopsidane.
Subtribe 10. ENHYDRINAE
Stuessy (1973) transferred the genus to Coreopsidinae, which it resembles in habit and involucre details. In specimens showing colored resin ducts in the corolla lobes of the disc, the ducts are within the veins and do not join where they terminate at the tip of the throat. In the species I have sampled, there is no variation in the dense pubescence of the anther filaments or in the prominent thickenings along the vertical walls of the endothecial cells.
Subtribe 11. MONTANOINAE
Montanoinae's position in the current subtribal sequence is based primarily on the fact that the bisexual disc flowers have uncompressed, epappose achenes. Rojasianthe of the latter subtribe is of interest because it is the only other genus in the tribe with paleae that are markedly enlarged with age, but the presence of a pappus and the specialized form of disc corolla make the close relationship with Montanoa doubtful. The achenes of M o n - tanoa are more like those of Melampodiinae and Milleriinae, but the scarcely lobed sterile rays and the hermaphroditic disk flowers preclude inclusion of the genus in either group.
Subtribe 12. RUDBECKIINAE
Previous treatments have linked Montanoa with various other genera with sterile ray flowers, genera here in Helianthinae and Ecliptinae. Ripe spermatozoa of Rqasianthe were not seen, but there appeared to be no furrows in the wall. Echinacea is apparently not very close to the Rudbeckiinae and agrees in all essential details with the present interpretation of the Ecliptinae.
Subtribe 13. ZALUZANIINAE
Subtribe 14. HEPTANTHINAE
Ray achenes polygonal or curly, with brownish carbonization, not striate, sometimes winged; seed coat cells rather subquadrate, strongly ornamented; pappus of radially arranged squamels or arists; disk achenes linear, inoperable. The Heptanthinae are treated here as a natural group of two Cuban genera that differ in their small erect rosuliform habitus, sterile disk flowers, and the distinct, radially arranged pappus of rayed urethras. The genera differ from each other in a number of characteristics, including the shape of the leaf blades, which are heart-shaped with palmate veins in Heptanthus and elliptic with subpinnate veins in Tetraperone.
Subtribe 15. ECLIPTINAE
Many members have filaments along the veins of the discal corollas, a feature not seen elsewhere in Heliantheae. In the latter case, the glands appear both on the branches and on the upper part of the stem. The Encelia group was often placed close to members of the Helianthus group, especially Simsia, and the group remained within the Helianthus group.
Subtribe 16. HELIANTHINAE
All members have sterile rays, single stigmatic surface on style branches of disc flowers, extraveinal canals on styles, and striae on achene walls. Most members of Helianthinae differ from Galinsoginae in having single resin ducts along the throat veins of the disc corolla. The termination of the canal is usually abrupt and no canal is visible on the axis of the style.
Subtribe 17. NEUROLAENINAE
Hehomeris of the western United States and northern Mexico is hardly distinguished in any other way. Both subtribes share prismatic achenes, a radially arranged pappus, colored resin, a position of the resin tube within the veins in the shaft of the style, and usually a subimbricate involucre. Tetrachyron is kept separate from Calea, but i t is similar to that genus in the position of the style.
Subtribe 18. GALINSOCINAE
The only member of Neurolaeninae included was Calea, but the genus Neurolaena itself was excluded from the tribe. The prismatic achenes, radially symmetrical pappus, and often subimbricate involucre of Galinsoginae are consistent with conditions considered primitive in the tribe in this study. Limited observations of the surfaces of the ovule indicate potentially useful variations in the rootstock.
Subtribe 19. DIMERESIINAE
T he carpopodium of the rayed achenes, the double or even triple resin canals along the veins of the disk corolla throat, and the tubes within the veins of the style seem to confirm this placement. T he trifid paleae of the receptacle and the pappus confined to the disk pinnae find their equals in several other genera of the subtribe. The anthers are actually barely longer at the base than many other members of the Heliantheae, and the tips are not caudate.
Subtribe 20. COREOPSIDINAE
In the study by Crawford and Stuessy (1981) , all 25 genera tested that were placed in the Coreopsidane by both Stuessy (1977) and the present treatment were found to have anthochlore. Aromatic end groups are present in the acetylenes of many Coreopsidinae, formed at both ends of the molecule. Such aromatic acetylenes are found elsewhere in the tribe only in Heliopsis of the Ecliptinae.
Subtribe 21. FITCHIINAE
The mutual occurrence of anthochlors in Coreopsidinae and in Lasthenia parallels the apparent restriction of thiophene forms of the en-tetrayne-en-polyacetylenes to the same two groups in Heliantheae. Coreopsidinae appears to have a greater diversity of polyacetylenes than any other subtribe of Heliantheae. The specialized homogamous ligulate heads of Fitchia resemble those of the Lactuceae (Cichorieae), and the genus was placed in this tribe by Bentham and Hooker (1873) and Hoffman.
COULTERELLINAE
Stebbins (1933) rejected the genus of the Lactuceae and proposed a position in the Mutisieae or Heliantheae. T he proper relationship and clear subtribe status of Fitchia was established by Carlquist (1957) in his detailed taxonomic and anatomical study of the genus. T he involucres of Coulterella can be interpreted to have a slight glandular development at the top, but lack the distinct type of gland characteristic of the Pectidinae.
Subtribe 23. PECTIDINAE
Colored resin has been seen in small amounts in the disc corollas of Adenopappus and Nicolletia. Papillae are restricted to the margins of the lobes in various genera including iVicolletia and in the Peruvian species Dyssodia lopez-mirandae Cabrera which Strother (1969) would place in Schizotrichia. Reversing an earlier judgment (Robinson, 1973), Vilobia is distinguished from Tagetes on the basis of a laciniate pappus type that is unmatched among the many variations found in the latter genus.
Subtribe 24. FLAVERIINAE
The substrain is widely distributed in the drier and more open habitats of tropical and subtropical America. Regarding pappus variation in the subtribe, it is notable that an obsolete pappus occurs in 2 species of Flaveria (Powell, 1978). Tri-thiophenes are present and are known elsewhere in the tribe Pectidinae, a Gaillardia and in Eclipta.
Subtribe 25. VARILLINAE
The current concept of Flaveriinae follows that of Turner and Johnston (1961) and Turner and Powell (1957). These are a factor in excluding all other genera from the current subtribe concept. This type finds its best expression in the reddish canals in the crown nozzles of the genus Varilla.
CLAPPIINAE
The pattern of a prominent resin canal between two veins is one of three possible basic types, but it is the least common type in the tribe. In geographic distribution, xeric nature, shrubby habit, pale stems with alternate leaves, and brittle condition of dried plants, the Clappiinae closely resemble at least one of the species of Varillas. In such a grouping, the epaleal condition appears to be more primitive, and Varilla's paleae would be a revived structure in the form of homologous involucral bracts.
JAUMEINAE
Anatomical study shows additional similarities in the costate achenes, the elongated, nonpapillose cells of the disc corolla lobes, and the tendency for a reddish resin tube between paired veins in the disc corolla throat. T he superficial resemblance of the resinous lines of Cluppiu to the secretory cavities of the Pectidinae is misleading, and the latter subtribe is more remote in proportion. T he closest relationship of the Jaumeinae is probably to subtribes such as the Clappiinae and Flaveriinae, which also have multi-costate achenes.
Subtribe 28. MADIINAE
Members of the Madiinae are commonly known as tar weeds, evidently from the general occurrence of abundant pectic material in the leaves. T he achene walls of the Madiinae appear to be basically striated, and the cells of the seed coat basically unornamented, but interesting variations occur in both structures. However, no close relatives are apparent among the paleaceous Heliantheae, and in many respects the Madiinae appear to be members of the epalaceous series.
HYMENOPAPPINAE
The achenes of the Hymenopappinae also tend to be distinguished by the triangular or quadrangular shape and stipitate bases. The base of the fruit is not stipitate, and the cells of the seed coat are slightly deformed without any decoration on the walls. The plant is, by all indications, a member of the epalaceous series, but the anther thecae are distinctly black.
Subtribe 30. LYCAPSINAE
T he Hymenopappinae differs from the rest of the series by the distinctly carbonized achene walls that lack striations. T he genus was resurrected and removed from the Eupatorieae by Johnston (1935) who noted the radiating condition of the heads. As noted by Johnston (1935), the genus is suggestive of the helenioids, although it has paleae.
PERITYLINAE
T he taxonomic history was recently revised in a survey of the subtribe by Powell and Turner (1974). Eatonella and Oxypappus with X = 10 appear to be members of the subtribe Baeriinae, and Hulsea with X = 19 is a member of the Chaenactidinae. T he endothesis, cells of all but one genus of the substem are characterized by 2 or 3 thickenings on the transverse walls.
Subtribe 32. BAERIINAE
Most genera of the Baeriinae are further distinguished from the related Peritylinae by the broader bracts. Oxypappus is apparently related to this subtribe by the shape of the leaves and the achenes. Several members of the subtribe examined show weakly to strongly developed conical receptacles.
Subtribe 33. CHAENACTIDINAE
Achenes of the group usually show a regular pattern of nodular thickenings in carbonization at full maturity. Genera in the group have carbonization of the achene wall without any additional pattern of nodular thickenings. Psathyrotopsis is a member of the Chaenactidinae, while the typical Psathyrotes is placed in the Gaillardiinae near Trichoptilium.
GAILLARDIINAE
Helenieae is now generally accepted as an integral part of the Heliantheae (Stuessy, 1977; Turner and Powell, 1977). Both genera have broad cells covering the surface of the achene with dark granular inclusions in the lumina, unlike other genera of the subtribe. Helenanolides have also been reported from Arnica of the closely related subtribe Chaenactidinae and from two genera, Geigeria Griessel.
Subtribe 35. MARSHALLIINAE
Careful redefinition of the boundaries between the two tribes may ultimately indicate that the Marshalliinae can best be accommodated in the Inuleae. Relationships of 8 tribes of Compositae as suggested by plastocyanin amino acid sequence data. Evolution of compound capitulum types in the light of insect-flower interaction.
