Update: 2012-12-31 06:22 PM +0630



04.2 Criteria of classification


by George H. M. Lawrence, Professor of Botany at the Bailey Hortorium, Cornell University, 1951

Photo-copied by Maung Kan Tun from the original text.
Edited and set in HTML by U Kyaw Tun (M.S., I.P.S.T., USA), Tun Institute of Learning (TIL) and the staff of TIL. Not for sale. No copyright. Free for everyone. Prepared for students and staff of TIL  Computing and Language Center, Yangon, MYANMAR :  http://www.tuninst.net , http://www.softguide.net.mm , www.romabama.blogspot.com

index.htm | Top

Contents of this page

Principles of Taxonomy
Criteria of classification
- Morphological criteria p056
    Vegetative vs reproductive structures p057
    Ligneous vs. herbaceous characters p058
    Inflorescences p059
    Evolution of inflorescences p060
    Fig. 3. Schematic diagrams of hypothetical evolution
    of inflorescence types 063
Foot notes
UKT notes
  Vegetative structures

Contents of this page

04.2 Criteria of classification
Morphological criteria

[{cont. p056}]

All taxonomists are agreed that the differences between plants, and the similarities that plants may possess in common, are measurable to a large degree by the morphological characters of those plants. A morphological character, to a taxonomist, is one inherent in or manifested by

Lawrence 057

a structural component of the plant. In general it is concerned with the organography of the plant, with characters that usually are discernible with the aid of no more than a good hand lens. The more recent recognitions by taxonomists of the significance of the related sciences of anatomy, cytology, and comparative morphology have resulted in an expansion of the concept as to what is a morphological character and to the extreme that some taxonomists now consider the number of chromosomes to represent as significant a morphological character as that provided by the number of stamens. The value of the morphological character is measured by its constancy. The more constant the character, the greater is the reliability that can be placed upon it. In this regard it must be remembered that a character that is reasonably constant among the members of one group may be a weak character among members of another group. For example, among the families of some dicotyledonous plants the character of ovary position (inferior vs. superior) is adequate to separate families and orders, but within some monocotyledonous plants the two types of ovary position are to be encountered within a family (e.g., Bromeliaceae), or even within a single genus (e.g., Bomarœa).

Contents of this page

Vegetative vs. reproductive structures.

The vegetative structures of plants (i.e., those of leaves, stems, buds, and habit of growth) are readily apparent to the eye and have an appeal to the beginning taxonomist. Their observation generally does not require the use of a microscope or high-powered lens nor the employment of razor blade or dissecting needles. However, their use as taxonomic criteria is limited for two basic reason: the total number of vegetative characters available is few as compared with the number of species of vascular plants, and too often the vegetative characters are not particularly constant.

The number of species of woody plants of any temperate flora is usually fewer than the number of species of herbaceous plants of the same flora. For this reason, most genera, and sometimes the species, of these woody plants may be separated largely and almost exclusively by vegetative characters. Keys to woody plants are popular for this reason. However, even among the genera of woody plants, vegetative characters lack the necessary variability of type, stability, and constancy necessary for the identification of species. Examples are abundant in such temperate genera as Crataegus, Prunus, Quercus, Rhododendron, Salix, and Vaccinium. As a rule, the greater the number of species in any genus of woody plants the more difficult is their identification by means of vegetative characters.

Lawrence 058

In the preparation of keys13 (fn058-13) to woody plants it frequently is desirable to use vegetative characters to as great an extent as practicable. It has been found from experience, in dealing with plants of temperate regions, that certain vegetative characters are generally more constant than are others. (In certain genera and families, however, there are notable exceptions.) Those characters known to be more reliable than others are listed below in sequence of approximate descending constancy:

Leaves needle- or scalelike vs. leaves broad and laminate
Venation parallel vs. venation reticulate (netted)
Leaves compound vs. leaves simple
Plants erect trees or shrubs vs. plants vinelike
Stipules present vs. stipules absent
Leaves lobed or divided vs. leaves neither lobes nor divided
Leaves opposite or whorled vs. leaves alternate
Leaf margins entire vs. leaf margins variously toothed
Leaves persistent and coriaceous vs. leaves deciduous
Spines or prickles present vs. spines or prickles absent
Leaf base tapering (cuneate) vs. leaf base broad (cordate, obtuse, truncate) (UKT: see leaf shapes )
Leaves with 1 main mid-vein vs. leaves with 3 primary veins

Succeeding the above would be characters of leaf form, leaf apex, the leaves petioled or sessile, and pith characters of the twig or stem. Characters of vesture (pubescence), surface, coloration, milky vs. watery sap, and sizes of plant or leaves are generally useful only in the separation of species or their variants.

See UKT note on vegetative structures adapted from on Trees Commonly Cultivated in Southeast Asia, an illustrated field-guide, 2nd ed., FAO, 1999, pp.230

Reproductive structures are those associated with the flower and fruit. The number of organs in these structures, and the number of components of these organs, are many times that available from vegetative structures. For this reason the number of mathematical combinations and recombinations of reproductive characters is infinitely greater and ordinarily exceeds the number of species (or smaller units) to be differentiated. In addition to being more abundant, reproductive characters in general are much more nearly constant than are vegetative characters. The reliability of any single reproductive character selected for taxonomic use must be determined individually, however, for each particular group of plants.

Contents of this page

Ligneous vs. herbaceous characters

In recent decades increased emphasis has been placed on the phyletic significance of the woody as opposed to the herbaceous character of plants. It is believed by many botanists that among the higher vascular plants, those of woody habit were the antecedents of those of herbaccous habit. Hutchinson, in the latest version of his classification (1948), has made this hypothesis the

Lawrence 059

basis of his two classes (Lignosae and Herbaceae) of the Dicotyledoneae. In doing so he warned

. . . against taking the view that he has reverted to the old practice of putting all the woody plants into one group and all the herbaceous plants into another. . . . what has been done is to put all those families which are predominately woody, and the  more primitive genera of which are woody, into the woody phylum, and conversely the same has been done with the herbaceous families.

In other words, the woody division consists of trees and shrubs plus herbs which are related to trees and shrubs, and the herbaceous division consists of herbs plus some clearly related woody plants.

In addition to the significance placed on these criteria by Hutchinson, the more elemental distinction between woody and herbaceous plants has practical value in artificially separating the minor units of classification. In many cases, the demarcation between the two criteria is not sharp. Some plants that appear to be herbaceuos are actually woody in that their stems in addition to being woody in texture are of perennial duration (e.g., Gaultheria procumbens, Epigaea repens, or Vinca minor). Other plants are basally woody but have stems that for the most part are herbaceous. Such plants are often referred to as being suffrutescent in character (e.g., Pachysandra terminalis, Arabis caucasica, or Aralia hispida).

Contents of this page


An inflorescence conventionally is considered to be the arrangement of flowers on the floral axis. By Linnaeus (1751), it was said to be a mode of flowering; that is, he considered it to be a phenomenon or function and not a structural component of the plant. The present-day classification of inflorescences, although utilizing to an extensive degree the terminology of Linnaeus, is based on the relatively unscientific observations of Link (1807) as modified by Roeper (1826)14 (fn059-14) These botanists considered the inflorescence to be a structural entity of the plant, subject to subdivision into definite classes and types for use as taxonomic criteria. The basic classification of the inflorescence by Roeper (into a series of determinate and indeterminate types) is, with some amplifications and modifications, that used today in taxonomic and less technical descriptions of flowering plants.15 (fn059-15) As pointed out by Rickett (p.189), inflorescences are branch systems and are "all related

Lawrence 060

in the sense of being derivable from one underlying type." It is conceded generally that the primitive inflorescence or its antecedent was a system of branches each of which was terminated by a flower. Every individual flower is a terminal unit. It is borne on the end of an axis, and it matters not whether that axis conventionally be termed a  peduncle, pedicel, receptacle, torus, or stem. The structure represented by any one of these terms is technically a floral axis. It is important to an understanding of the use of inflorescences in the classification of these plants to know what type of inflorescence is primitive, and of the possible developmental sequence of other types from the primitive. Unfortunately there is yet insufficient knowledge concerning the phyletic  relationships of the families of flowering plants to permit more than speculation on the origins of inflorescences. However, speculations based on what are believed to be the best available data do provide rational approaches to a solution of the problem.

Contents of this page

Evolution of inflorescences

Three theories have been proposed to account for the evolution of present-day inflorescences from the presumed primitive condition of a branch system of terminal flowers. The first postulates the primitive type to have been a panicle, the second that it was the single terminal flower, and the third that it was a dichasium. The first theory has been held by many investigators, including Nägeli (1883), Celakovskż (1892), and Pilger (1922). By acceptance of the  panicle as primitive, it is easy to explain the derivation of all other types from it  as was done in part by Goebel (1931). Pilger supported his views in defense of the theory with analogies which he conceded do not exist among living woody angiosperms. The second theory, supported by Parkin (1914) proposes the development of multiflowered inflorescences from the solitary flower by addition of lateral flowers (and branches of flowers) by axillary  growth. Any consideration of this theory must take into account the generally accepted view, discussed later in his chapter, that branch systems are a fundamental part of the plant organization, and that the flower is probably made up of highly modified branch systems. Acceptance of the belief that foliar organs are highly modified branch systems adds logic to a theory that the primitive inflorescence also was a branch system and that most existing examples of solitary flowers represent examples of reduction and suppression. The third theory, that the dichasium (Fig. 3)* may represent the primitive type, has been accepted by Rickett (1944).

*UKT: For reader's convenience this place is bookmarked as fig3call1b.

In its simplest form the primitive dichasium consisted of three flowers disposed in a cluster formed on a single peduncle by a dichotomous branching immediately beneath a terminal flower (Fig. 3a). From this

Lawrence 061

simple dichasium there may be developed a more complex dichasium that is formed by "a repetition of the same apparent dichotomy in each lateral branch." In nature, the dichasium may be simple or compound, ample or restricted, and it would be hazardous to speculate as to which form preceded the other in ancestral types. There is no clear paleobotanical evidence to support the view that the dichasium was ancestral to the panicle, neither can the latter be established as the more primitive. Either one may have been derived from the other. There is some fossil evidence that in a few rare instances the solitary flower may represent a primitive situation. It  has been shown (Florin, 1948) that primitive and now extinct conifers possessed a paniculate inflorescence, a fact that should not, however, be construed as evidence indicative of gymnospermous ancestry of the angiosperms. There is a theory held by many botanists that the Pteridospermae (seed plants with fernlike foliage) provide the ancestral group best possessing angiospermlike morphology; in this taxon the ovules, borne on complex fernlike leaves, suggest a paniculate arrangement. Irrespective of the relative primitiveness of the panicle and the dichasium, the views of Rickett on the possible evolution of inflorescence types, as outlined below, are most helpful and provide a logical understanding of possible relationships.

The dichasium was postulated by Rickett to have given rise, by only   slight structural modifications, to inflorescence types such as bostryches, cincinni, verticels, cymes, helicoid cymes, scorpioid cymes, and by apparent elimination of internodes it may have been an antecedent to types such as the umbel, capitulum [UKT: head], and corymb (Fig. 3). The obvious structural organization of  complex and congested inflorescences indicates that  many of them may have developed from various combinations and aggregations of dichasia. In this regard, it is now believed that the seemingly simple raceme  and spike   represent types that are reduced, along one of two or more evolutionary lines, from more complex ancestral types of dichasia or panicles. That is, the raceme or the spike may have been evolved through steps involving the reduction of individual dichasia to as little as a single flower accompanied or not by similar reduction of leaves to subtending bracts. It has been established that, in many cases, the presumed distinction of an inflorescence being determinate (i.e., centripetal) or indeterminate  (centrifugal) is without validity and that ultimately in a new classification of inflorescence types, this criterion will have less value than is currently accorded it. In certain inflorescence types it is a valid characteristic, while in others it is unreliable. For example, the umbel of members of the Umbelliferae is an indeterminate

Lawrence 062

type of inflorescence whose outer flowers open before the inner ones. However, in many of the Amaryllidaceae (as in Allium 16 (fn062-16)) the umbel is a determinate inflorescence as indicated by the central flower opening prior to the outer flowers. Structurally, the inflorescence is an umbel in either family, but phyletically the two kinds of umbel are unrelated and represent reductions by two different evolutionary lines (Fig. 3).

A derivative and advanced group of  inflorescence types over those considered to have more direct affinity with the dichasium is that known as the monochasium. A monochasium is an inflorescence resulting from a dichasium in which one branch of each dichotomy continues to develop while the other branch is suppressed completely (e.g., aborts). The result is a sympodial axis composed of series of superposed axis and is sometimes referred to as a sympodial dichotomy (Fig.3). This type of inflorescence is phyletically a complex inflorescence, since it is made up of a series of branch systems. It is the "scorpioid cyme" of botanical floras and manuals and is characteristic of many boraginaceous genera (Myosotis, Heliotropium, Symphytum) as well as of representatives of the   Polemoniaceae and Hydrophyllaceae.17 (fn062-17) [ {hsing-nha-maung:kri:} Heliotropium indicum L. , Family: Boraginaceae.] In these taxa its character and origin are usually obvious by macroscopic inspection. However, reduction often has advanced to a degree where the monochasium is not as apparent in all cases as in the so-called  scorpioid cyme. That is, in some groups, the sympodial false axis(actually composed of determinate units) may be so refined in appearance as to seem to be a simple and true axis of indeterminate growth; such an inflorescence is the raceme of Claytonia, an infloresence that morphologically is as much a helicoid cyme as is that of Myosotis. In other genera, the monochasium has been reduced to an umbel as in the florist's geranum (Pelargonium). The umbell of Allium, mentioned above, is of similar origin. From this brief review of possible origins of inflorescences, it may be speculated that either the panicle or the dichasium (rarely the solitary flower) is the primitive type of inflorescence, and that from them have been evolved all other types of inflorescences. That is, the solitary flower, the panicle, and the dichasium all may be primitive types.18(fn062-18) By this speculation, 

Lawrence 063

Fig. 3. Schematic diagrams of hypothetical evolution of inflorescence types

Fig. 3. Schematic diagrams of hypothetical evolution of inflorescence types:
a. compound diachasium; b. helicoid cyme; c. cincinnus;
d. scorpioid cyme; e. thyrse; f. panicle;
g. compound corymb; h. simple corymb; i. raceme;
j. indeterminate umbel; k. spike; l. indeterminated head;
m. cyme; n. determinate umbel; o. determinate head;
p. verticillate inflorescence; q. simple dichasium.
Go back to fig3call1b/ fig3call2b/ fig3call3b/ fig3call4b

Lawrence 064

it is indicative that usually the conventional raceme is more primitive than the spike but that both are advanced over the panicle, that while the panicle may have been primitive, some present-day examples of the panicle may have been derived from a thyrse, and that the numerous cymose types of inflorescences are perhaps more primitive than most of those mentioned above. Accordingly, the modern or present-day inflorescence represented by a solitary flower in almost all plants is a product of reduction, a reduction that may have resulted from any one of many multiflowered inflorescence type, whose origin may be indicated only by comparative studies of infloresences of related taxa.

The significance of the inflorescence as a taxonomic criterion has been established. In some families it is of significance in recognizing entire families (Umbelliferae, Cornaceae, Compositae, Labiatae, Gramineae, Amaryllidaceae [sensu Hutchinson]) and in others in recognizing tribes or genera (Galium, Smilax, Malus, Philadelphus). The consideration of phyletic significance of the inflorescence cannot be evaluated apart from other criteria, but in dealing with minor units of flowering plants often it may be helpful in establishing their position in a particular schema or classification.

Contents of this page


fn058-13. Cf. pp. 225- 228 for discussion of keys, their preparation, and use.
Go back fn058-13b

fn059-14. See Rickett, H. W. (1944), for a detailed analysis of the subject of inflorescence classification, and from whose review much of the data for this section of the chapter were obtained. See also Croizat (1944) for views treating the inflorescence as an aggregate of vegetative and floriferous axes.
Go back fn059-14b

fn059-15. A determinate inflorescence, simply expressed, is one in which the terminal flower of the axis opens first with the remaining flowers of that axis opening successively from top to bottom. An indeterminate inflorescence is one in which the lowermost flowers of an axis open first and the terminal flowers open last, with the successive opening of flowers accompanied by an elongation of the axis.
Go back fn059-15b

fn062-16.  For discussion of the family to which Allium belongs (Liliaceae vs. Amaryllidaceae), cf. Part II, p. 415.
Go back fn062-16b

fn062-17. Note from study of Fig. 3, That the so-called scorpioid cyme here referred to in quoted is properly a helicoid  cyme (helicos, mening , snaillike) and that the true scorpioid cyme is an inflorescence type not commonly encountered.
Go back fn062-17b

fn062-18. The pteridosperm leaves beat ovules in ways that would provide a background and historical basis for this: the ovule terminal and solitary on huge fronds; the ovules terminal on few or all pinnules; or the ovules marginal and terminal throughout the
Go back fn062-18b

Contents of this page

UKT notes:

stamen (def.)

stamen n. pl. stamens or stamina 1. The pollen-producing reproductive organ of a flower, usually consisting of a filament and an anther. [Latin stāmen thread] -- AHTD

Go back stamen-def-b


capitulum n. pl. capitula 1. Botany See flower head. -- AHTD
head n. 20. Botany A flower head. -- AHTD
flower head n. 1. Botany A dense, short, compact cluster of sessile flowers, as of composite plants or clover. Also Called capitulum . 2. A very dense grouping of flower buds, as in broccoli and cauliflower. -- AHTD

Go back capitulum-b

Contents of this page

sympodial axis

sympodium n. pl. sympodia Botany 1. A primary axis that develops from a series of short lateral branches and often has a zigzag or irregular form, as in orchids of the genus Cattleya. Also Called pseudaxis . -- AHTD

Go back sympod-axis-b

Contents of this page


stipule n. Botany 1. One of the usually small, paired appendages at the base of a leafstalk in certain plants, such as roses and beans. [New Latin stipula from Latin stalk] -- AHTD

Go back stipule-b

Contents of this page


coriaceous adj. 1. Of or like leather, especially in texture. [From Late Latin cori³ceus from Latin corium leather] -- AHTD

Go back coriaceous-b

Contents of this page

Vegetative structures

For identification of a plant or differentiation between two similar plants in field-work, vegetative structure are more important than reproductive structures: the primary reason being vegetative structures can be observed all year round whereas reproductive structures can be observed only for a short period. Among the vegetative structures, the character of the leaves seems to be the most important :

1. Arrangement of leaves:
opposite, alternate, whorled

2. Leaf structures:
simple, palmately compound, even pinnately compound, odd-pinnately compound, bipinnately compound. (pinnate also pinnated adj. 1. Resembling a feather; having parts or branches arranged on each side of a common axis: a polyp with a pinnate form; pinnate leaves. [Latin pinn³tus feathered from pinna feather; See pet- in Indo-European Roots.] -- AHTD)

3. Leaf shapes:
needle, scalelike, linear, lanceolate, oblong, elliptical, spatulate, ovate, orbicular, reniform, cordate, obovate, cuneate, oblanceolate, falcate
(Go back leaf-shape-b)

4. Leaf margins:

entire, sinuate, serrate, dentate, lobed, clefted, crenate, incised, parted, divided


5. Leaf tips:
acuminate, acute, obtuse, rounded, truncate, emarginate

6. Leaf bases:
cuneate, acute, obtuse, rounded, truncate, auricula (auricle n. 2. Biology An earlobe-shaped part, process, or appendage, especially at the base of an organ. -- AHTD).(Go back leaf-base-b)

7. Leaf venation:
pinnate, arcuate, parallel, palmate

Go back to veg-structures-b


Contents of this page


determinate adj. 4. Botany a. Terminating in a flower and blooming in a sequence beginning with the uppermost or central flower; cymose: a determinate inflorescence. b. Not continuing indefinitely at the tip of an axis: determinate growth. -- AHTD

Go back determin-b

Contents of this page

End of TIL file