TRENDS IN SYSTEMATIC BOTANY
61
at the generic level is demonstrated in a number of papers by Copeland
(1939, 1941, 1943, 1947, 1954b) and in one by Thomson (1942).
VI. POLLEN MORPHOLOGY
The form of the pollen grain offers another little-used morphological
character of taxonomic importance. It helps to mark out the general
broad lines of relationship, as for example, most dicotyledons have the
very distinctive tricolpate type of pollen, or grains apparently derived
from this type. Such grains are not known in any other group of seed
plants and are believed to be specialized. In a number of the woody
families of the Ranales, however, one finds monocolpate pollen, a less
specialized type, which is of common occurrence in the monocots and is
characteristic of many seed ferns, Bennettitales, Cycadales, and Ginkgoales (Erdtman, 1952).
Pollen morphology has proved very useful as an additional character
in certain cases on anomalous genera of dubious family affinity that
seem to fit with equal logic into distantly related families (Dahl, 1955).
Even within genera a careful study of pollen morphology may be
helpful. Hedberg (1946) discovered ten well-defined. types of pollen
grains in the genus Polygonum s. lat., each of which occurred in only
one definite part of the genus. Some species of debated relationship
could easily be referred to their proper groups by using these data, and,
on the basis of general morphology, pollen morphology, and chromosome
numbers, Hedberg suggested a division of Polygonum s. lat. into seven
genera. The generic limits of this complex have been changed many
times by various workers, and Hedberg's conclusions may be the nearest
correct. It should be borne in mind, however, that pollen morphology
can take us far astray, as can any other single character when used to
determine relationships. This is well illustrated in the umbelliferous
genus Sanícula, in which Bell (1954) has shown that both spherical and
ellipsoidal pollen grains of amply distinct morphology occur in equal
numbers and are equally functional in anthers on the same plant.
In an exhaustive study of some 350 species representing all sections of
Centaurea in the broadest sense, Wagenitz (1955) found eight distinct
pollen types represented. In Centaurea in the restricted sense, mostly one
type of pollen was found. Discovery of pollen of another type was cause
for rejecting the species from the genus thus restricted, and such action
was substantiated by other evidence. Wagenitz believes that in this complex too great stress has been laid on pappus and involucral bract characters alone, and that the pollen characters taken in conjunction with these
offer an improved classification.
61
at the generic level is demonstrated in a number of papers by Copeland
(1939, 1941, 1943, 1947, 1954b) and in one by Thomson (1942).
VI. POLLEN MORPHOLOGY
The form of the pollen grain offers another little-used morphological
character of taxonomic importance. It helps to mark out the general
broad lines of relationship, as for example, most dicotyledons have the
very distinctive tricolpate type of pollen, or grains apparently derived
from this type. Such grains are not known in any other group of seed
plants and are believed to be specialized. In a number of the woody
families of the Ranales, however, one finds monocolpate pollen, a less
specialized type, which is of common occurrence in the monocots and is
characteristic of many seed ferns, Bennettitales, Cycadales, and Ginkgoales (Erdtman, 1952).
Pollen morphology has proved very useful as an additional character
in certain cases on anomalous genera of dubious family affinity that
seem to fit with equal logic into distantly related families (Dahl, 1955).
Even within genera a careful study of pollen morphology may be
helpful. Hedberg (1946) discovered ten well-defined. types of pollen
grains in the genus Polygonum s. lat., each of which occurred in only
one definite part of the genus. Some species of debated relationship
could easily be referred to their proper groups by using these data, and,
on the basis of general morphology, pollen morphology, and chromosome
numbers, Hedberg suggested a division of Polygonum s. lat. into seven
genera. The generic limits of this complex have been changed many
times by various workers, and Hedberg's conclusions may be the nearest
correct. It should be borne in mind, however, that pollen morphology
can take us far astray, as can any other single character when used to
determine relationships. This is well illustrated in the umbelliferous
genus Sanícula, in which Bell (1954) has shown that both spherical and
ellipsoidal pollen grains of amply distinct morphology occur in equal
numbers and are equally functional in anthers on the same plant.
In an exhaustive study of some 350 species representing all sections of
Centaurea in the broadest sense, Wagenitz (1955) found eight distinct
pollen types represented. In Centaurea in the restricted sense, mostly one
type of pollen was found. Discovery of pollen of another type was cause
for rejecting the species from the genus thus restricted, and such action
was substantiated by other evidence. Wagenitz believes that in this complex too great stress has been laid on pappus and involucral bract characters alone, and that the pollen characters taken in conjunction with these
offer an improved classification.