Hairs of 412 taxa representing 66 genera of New World Melastomataceae were examined by SEM. Representatives for each hair type are illustrated by SEM micrographs. Epicuticular waxes are illustrated. Unusual “stomatal” apertures on intersepalar hairs in the genera Pterolepis and Tibouchina were found. Parallel evolution of hairs in different genera and subgeneric divisions has occurred frequently. Hair diversity probably is the result of evolution for protection against arthropod predators.

In Part I, all known taxa of Begoniaceae (exclusive of hybrids and cultivars) are initially keyed into 34 subkeys and are subsequently keyed to all recognized species, with each species illustrated from a photo of the holotype or type specimen. Part II lists, in alphabetical order, all verified species and varietal epithets, including complete synonymy, literature citations, and geographical distribution. All the recognized species are cross referenced to the subkeys in Part 1 by the illustration numbers.

Despite substantial plant collecting in Bolivia, there is no general reference that summarizes information about collectors. Here, we list individuals known to have collected plants in Bolivia and for whom there is published information concerning their lives or their Bolivian collections. For each, we include biographical information, dates of Bolivian collections, herbaria containing the Bolivian specimens, and pertinent references to publications by the collector, or about the collector, the specimens, or the expedition. We also include a section of general references which pertain to several collectors; these are cross-referenced under the discussion of each collector. The type of information available from each general reference is briefly described.

Detailed descriptions and drawings of caryopses are presented for 35 species representing 18 genera in the Triticeae. We studied the morphological variation in non-embryo characters involving the length, shape, compression, lateral ridges, sulcus development, coronas, and apical hairs of the caryopsis. Based on shared characters of the caryopses, we recognize three groups of genera as two major subtribes and one monogeneric subtribe. The Hordeinae includes Agropyron, Crithopsis, Elymus, Elytrigia, Eremopyrum, Hordelymus, Hordeum, Leymus, Psathyrostachys, and Taeniatherum; Triticinae includes Aegilops, Amblyopyrum, Dasypyrum, Heteranthelium, Secale, and Triticum; and Henrardiinae includes Henrardia. Similarities among the caryopses of 14 species of Elymus, Elytrigia, Leymus, and Psathyrostachys support a broad interpretation of Elymus. The caryopses of Brachypodium resemble those of the genera in the Hordeinae and to some extent Bromus.

Chromosome reports are provided for 145 populations, including first reports for 33 species and three genera, Garcilassa, Riencourtia, and Helianthopsis. Chromosome numbers are arranged according to Robinson's recently broadened concept of the Heliantheae, with citations for 212 of the ca. 265 genera and 32 of the 35 subtribes. Diverse elements, including the Ambrosieae, typical Heliantheae, most Helenieae, the Tegeteae, and genera such as Arnica from the Senecioneae, are seen to share a specialized cytological history involving polyploid ancestry. The authors disagree with one another regarding the point at which such polyploidy occurred and on whether subtribes lacking higher numbers, such as the Galinsoginae, share the polyploid ancestry. Numerous examples of aneuploid decrease, secondary polyploidy, and some secondary aneuploid decreases are cited. The Marshalliinae are considered remote from other subtribes and close to the Inuleae. Evidence from related tribes favors an ultimate base of X = 10 for the Heliantheae and at least the subfamily Asteroideae.

In the present study a number of species originally assigned to Panicum but not fitting well within its boundaries have been examined. Some of the characters that have been used in the past to distinguish genera in the tribe have been re-examined, including type of hilum, spikelet compression, relative length of the glumes, consistency of the upper anthecium and whether crested or not, and carriage of the spikelet relative to the axis. Further characters of taxonomic importance have been revealed by studies made with the scanning electron microscope (SEM) at higher magnifications than previously possible. The results of these studies indicate that several species should be removed from Panicum and transferred to other genera. Accordingly the following changes have been made: P. megastachyum Nees ex Trinius and P. albicomum Swallen & García Barriga have been transferred to Brachiaria, P. glutinosum Swartz and P. villaricense Mez to Homolepis, P. grandifolium Doell to Ichnanthus, P. killipii Hitchcock to Paspalum, and P. macranthum Trinius to Streptostachys. Panicum arnacites Trinius forms the basis of a new genus, Tatianyx. With these changes in content the descriptions of Homolepis and Streptostachys are emended, two new species—Brachiaria tatianae and Streptostachys ramosa—are described, and the unusual species P. aristellum Doell is characterized and a possible relationship discussed. A chart is presented that compares the characters of the genera covered in the present investigation, P. aristellum and the two new species are illustrated, and SEM photomicrographs of all taxa are included.

To some extent classification is subjective. Taxonomists differ in the relative importance they ascribe to particular characters and in the degree of difference between related taxa they deem sufficient to constitute family or ordinal rank. About 1000 dicot family names have been published. Those who have attempted an overview of the system at the family level and above in the last quarter century recognize between 274 and 455 dicot families in 39 to 82 orders. I accept 334 families and 59 orders. In Table 1 I give my ordinal allocation of the families and that of 11 recent authors to indicate where there is agreement and where there are differences to be resolved. I have also constructed a dendrogram to suggest relationships and degree of advancement of the orders.

I have written concise, uniform descriptions of all the families of dicots, emphasizing those characters that show trends between families or occur in more than one family. Each family is illustrated by analytical drawings of the flower, fruit, seed and usually inflorescence. Several species are usually used to show the range of major variation within families and trends toward related families.

Angiosperms probably arose from gymnosperms, so characters or character states universal in gymnosperms or considered primitive in them would also be considered primitive in angiosperms.

My approach to understanding evolutionary trends in characters is to relate them to the ecological factors that might be responsible for them by their selective action. The dicots probably originated under warm temperate conditions favorable for growth. A major evolutionary trend in them has been the gradual development of characters and character states enabling them to cope with dry and hot or cold conditions and colonize generally unfavorable regions.

A second major trend has been from wind pollination to progressively better adaptation for insect pollination. The primitive insect pollinated dicots often have flowers with numerous spirally arranged parts; plants having flowers with few, opposite or whorled parts are derived.

The floral organs are homologous with leaves. Like leaves the parts were initially separate. The connate and adnate conditions are derived.

General character states are primitive; specialized states are derived. In attempting to determine which primitive states are most primitive I considered their occurrence among the families. The fewer the families with a particular primative state, the more primitive the state. This is important in deciding whether a family is low or high on the family tree and the position within its particular order.

In accordance with the above rationale, I have constructed a table giving the primitive and derived states for about 100 characters. I also indicate the extent to which I consider the states reversible.

To determine a family's phylogeny, it must be compared with other families considered to be close to it. In general, the more characters and character states in common, particularly uncommon ones, the more likely are the subject families to be related. All parts of the plant and many characters should be considered. If a family has more than one state of a character, the state considered primitive for the particular taxon rather than the derived one should be considered in attempting to determine the extant family closest to its ancestor. A descendant has at least one more derived character or character state than its ancestor.

To some extent classification is subjective. Taxonomists differ in the relative importance they ascribe to particular characters and in the degree of difference between related taxa they deem sufficient to constitute family or ordinal rank. About 250 monocot family names have been published. Those who have attempted an overview of the system at the family level and above in the last quarter century recognize between 45 and 103 monocot families in 14 to 38 orders. I accept 57 families in 18 orders. In Table 1 I give my ordinal allocation of the families and that of 11 recent authors to indicate where there is agreement and where there are differences to be resolved. I have constructed a dendrogram to suggest relationships and degree of advancement of the orders.

I have written concise, uniform descriptions of all the families of monocots emphasizing those characters that show trends between families or occur in more than one family. Each family is illustrated by analytical drawings of the flower, fruit, seed, and usually inflorescence. Several species are usually used to show the range of major variation within families and trends toward related families.

Monocots and dicots have existed concurrently for most of their history, have been subjected to many of the same ecological pressures, and consequently show similar evolutionary trends. My approach to understanding evolutionary trends in characters is to relate them to the ecological factors that might be responsible for them by their selective action. The monocots probably originated under warm temperate or subtropical conditions favorable for growth. A major evolutionary trend in them has been the gradual development of characters and character states enabling them to cope with dry and hot or cold conditions and colonize generally unfavorable habitats.

A second major trend has been progressively greater specialization for insect pollination. The primitive monocots have flowers with numerous spirally arranged parts; those having flowers with few, opposite or whorled parts are derived.

The floral organs are homologous with leaves. Like leaves the parts were initially separate. The connate and adnate conditions are derived.

General character states are primitive; specialized states are derived. In attempting to determine which primitive states are most primitive I considered their occurrence among the families. The fewer the families with a particular primitive state, the more primitive the state. This is important in deciding whether a family is low or high on the family tree and the position within its particular order.

In accordance with the above rationale, I have constructed a table giving the primitive and derived states for about 85 characters. I also indicate the extent to which I consider the states reversible.

To determine a family's phylogeny, it must be compared with other families considered to be close to it. In general, the more characters and character states in common, particularly uncommon ones, the more likely are the subject families to be related. All parts of the plant and many characters should be considered. If a family has more than one state of a character, the state considered primitive for the particular family should be used in attempting to determine the extant family closest to its ancestor. A descendant has at least one more derived character or character state than its ancestor.

The xylem anatomy of 28 species of commercially and potentially commercial timbers of West Africa is described together with information pertaining to seasoning qualities, durability and working properties, as well as the uses of wood. A comprehensive discussion on the mechanical properties, establishing the methodologies for evaluating the potential utilization of these woods, has been included. Shrinkage and swelling in wood have always presented problems in the utilization of woods. A discussion relating to the differences among (a) moisture content change and shrinkage, (b) the effect of drying conditions on shrinkage and (c) the variation in shrinkage in different species is presented. To aid both beginning students and to refresh the minds of practicing wood technologists, a glossary of the principal terms used in describing the minute features of timbers has also been added.

Dalechampia, Haematostemon, Omphalea, Pera, Plukenetia, and Tragia (Euphorbiaceae, subfamily Acalyphoideae) are treated for the Guianas. Notes on Guianan Euphorbiaceae, subfamily Acalyphoideae, tribes Omphaleae, Pereae, and Plukenetieae, and ecological notes on each species are given. Also included is a key to the subfamilies of Euphorbiaceae and a key to the Guianan genera of subfamily Acalyphoideae.