De Candolle (1830) divided the genus Campanula into two large sections on basis of the presence or absence of calyx-appendages between the calyx-lobes. Boissier (1875) attached great value to the mode of dehiscence of the capsule, and divided the genus into two sections. None of the existing classifications seems to be a natural one. As cytological investigations and crossing experiments might give valuable information for a natural classification, it was decided to investigate: a. The classification of the species within the genus Campanula based on morphological, cytological, and genetic data. b. The variability of a number of species, based on cytological investigations and growing experiments carried out under uniform conditions. In Chapter I a survey is given of the most important literature on the classification of the genus Campanula. The cytological data, hitherto published, are listed in Chapter II. 77 species were studied cytologically, the chromosome numbers of plants of 729 different localities were counted. At the end of Chapter II some drawings of the somatic chromosomes of a number of species are given. The integration of cytological and morphological data is given in Chapter III. It appeared that, beside some rare chromosome numbers (2n = 24, 26, 28, 36, 56, 58), also some cytological series exist, each of which has its own basic number: x = 8, 10, 15, 17. Within each series the species usually show a great morphological resemblance. Also species studied by other authors show a combination of morphological and cytological characters corresponding with the correlations in the species which were studied by the present author. There are many reasons justifying the supposition that Sugiura, who reported many chromosome numbers, did not correctly identify the plants on which the chromosome count was based. In Chapter IV a survey of the results of the crossing experiments is given. The features pointing to relationship (dealt with in Chapter III) were tested by the crossing experiments. Some species with basal and apical dehiscence of the fruit are crossable. Hybrids were obtained from crosses between some species with and without calyx-appendages. Species belonging to different subsections of Fedorov’s system turned out to be crossable. In view of these facts the classifications given by de Candolle, Boissier and Fedorov cannot be regarded as natural. With the exception of species belonging to the x = 15- and the x = 17-series it was impossible to cross species belonging to different cytological series. From the selfpollination experiments the conclusion may be drawn that self-fertilization is a rarely occurring phenomenon in the genus Campanula. Most species investigated turned out to be self-sterile. Insect pollination is the rule, self-pollination the exception. As only 40-50 % of the total number of species of the genus Campanula have been investigated cytologically as well as morphologically, only a provisional division of the genus Campanula into a number of groups was given (Chapter V). These 7 groups are regarded as natural, but neither their interrelationship nor the relation of some of these groups to other genera of the family Campanulaceae is clear yet. At the end of Chapter V theories on the evolution of the chromosome numbers are discussed. The author gives an opinion differing from the one given by Böcher on the origin of some chromosome numbers. In Chapter VI a survey is given of the results of experimental cultivations of a great number of plants of 9 polymorphic species.