In the northern part of the Peruvian coastal region coconut palms were discovered which in addition to normal fruits produce larger numbers of bicarpellate fruits than have been previously recorded.
Deviations from the occurrence of 3 carpels in palm-fruits are discussed. In addition to the normal appearance of other carpel numbers in certain genera special attention is paid to unusual changes in their number: and particularly to decreases. It is shown -that the number of carpels is normally 3, but in some cases these do not develop equally and sometimes one or two remain rudimentary. In other cases they do not develop at all.
Comparison of the normal and bicarpellate coconuts shows a reduction of the third carpel in the latter, which has only one compressed locule, 2 'eyes' and 1 ridge. Anatomical analysis of the 'eyes' in normal and bicarpellate coconuts has resulted in the opinion that in all instances both the inner and outer integuments of each ovule coalesce '.with part of the inner fruit wall. A theory has been developed on the origin of the 'eyes'. A comparative study of the 'hard' and the 'soft eye' has resulted in a theory on their respective development and this theory explains why the 'soft' eye always corresponds to the functional carpel. When the non-functional ovules come under pressure from the strongly developing functional one, they gradually become compressed against the endocarp wall. The 'eyes' remain intact but become non-functional. This theory furthermore explains why in abnormal 2-or 3-seeded coconuts the corresponding 'eyes' are 'soft' or functional; because of their simultaneous development the 2 or 3 developing seeds establish a balance of pressure and consequently do not become compressed.
There is a marked similarity between the present author's findings in the coconut and studies on the development of the fruit in the African oil palm. Accordingly, the conclusion is drawn that the process of development of the 'eyes' in the fruits of both palms is very similar, if not identical. Following the use of the word 'operculum' to describe similar tissues in the fruit of the African oil palm, this term is also adopted for the coconut palm to describe the 2 layers which together close off the 'eye'.
The number of chromosomes from root tips of germinating normal and bicarpellate coconuts was found to be 32.
The variation among bicarpellate fruits is described.
Also a coconut is discussed in which the 'hard eyes' and the corresponding locules have nearly joined.
The analysis of the morphology of the inflorescence has shown that the triad of a female flower and two accompanying male flowers is a partial inflorescence of cymous character, a cincinnus. Pairs of male flowers represent a reduced stage of this cincinnus, as does the single male flower. The following abnormalities were observed: hermaphrodite flowers, abnormal male flowers and clusters of 2 male and 2 female flowers. The occurrence of this last abnormality has not yet been reported in normally branched inflorescences. A flower with a reduced number of perianth leaves was also found; its ovary was abnormal. Pairs of male flowers occurred in which the first flower was pedicellate, another abnormality which has not been described before. The analysis of these abnormalities confirms the theory on the morphology of inflorescences and flowers, which was initially only based on normal inflorescences and flowers.
The reduction in the number of carpels in the bicarpellate coconuts can be traced to the development of an abnormal syncarpous ovary. In the ripe bicarpellate fruits no remnants of the third carpel were found. Reduced numbers were registered, of floral envelopes, attached to both many of the bicarpellate fruits and to one female flower. Earlier studies on the morphogenesis of the flowers of the coconut palm, have served to determine the moment at which a flower, destined to become a bicarpellate fruit, is already abnormal. The moment at which the reduction of perianth leaves occurs is approx. 21 months prior to the harvesting of the bicarpellate fruits, while reduction in the number of carpels takes place about 18 1/2 months before harvesting.
Statistical analysis of the available data shows that the tendency to produce bicarpellate coconuts is most probably genetically determined.The main argumen for this theory is that no palms were observed which produced bicarpellate coconuts with yellow-ivory husk colour, although the percentage of coconut palms in the population studied, producing fruits with such husk colour was high. It is suggested that the oldest group of palms was composed of material of various origin. The remarkable yields of normal and bicarpellate fruits from two of the palms is also mentioned.
The influence of the environment, especially the weather has been studied. It is clear that the low winter temperatures favour the appearance of bicarpellate coconuts. Abnormally low temperatures result in high numbers of such fruits.