MALPIGHIACEAE : CORRELATIONS BETWEEN HABIT , FRUIT TYPE AND BASIC CHROMOSOME NUMBER

(Malpighiaceae: correlations between habit, fruit type and basic chromosome number). The family Malpighiaceae presents species with different habits, fruit types and cytological characters. Climbers are considered the most derived habit, followed, respectively, by the shrubby and arboreal ones. The present study examines the relationship between basic chromosome numbers and the derivation of climbing habit and fruit types in Malpighiaceae. A comparison of all the chromosome number reports for Malpighiaceae showed a predominance of chromosome numbers based on x=5 or 10 in the genera of sub-family Malpighioideae, mainly represented by climbers with winged fruits, whereas non-climbing species with non-winged fruits, which predominate in sub-family Byrsonimoideae, had counts based on x=6, which is considered the less derived basic number for the family. Based on such data, confirmed by statistic assays, and on the monophyletic origin of this family, we admit the hypothesis that morphological derivation of habit and fruit is correlated with chromosome basic number variation in the family Malpighiaceae.


Introduction
The family Malpighiaceae comprises about 71 genera and 1250 species.These species are mostly climbers, although trees and shrubs are also present (Robertson 1972;Judd et al. 1999).Some Malpighiaceae species are of economic interest, e.g.Barbados cherry (Malpighia emarginata), a rich source of vitamin C (Mezquita & Vigoa 2000).The genera Banisteriopsis and Diplopterys include some species with hallucinogenic compounds (Desmarchelier et al. 1996), the most famous of which is Banisteriopsis caapi, found in North of South America (Gates 1982).Species of the genera Banisteriopsis, Galphimia, Malpighia, Peixotoa and Stigmaphyllon are cultivated as ornamental plants (Lorenzi & Souza 1999).
Based on morphological characters and palynological data, Vogel (1990) characterized this family as a natural monophyletic group.It is taxonomically organized in two sub-families: Malpighioideae, with samaroid winged fruits that are wind dispersed, and Byrsonimoideae with schizocarpic, drupaceous fruits, which are dispersed by the water or animals (Anderson 1977;1993).Based partly on the anemocoric dispersion of their fruits and the climbing habit of its species, Anderson (1983) considered Malpighioideae the most derived sub-family.Although several habits are found in both subfamilies, lianas clearly predominate in Malpighioideae, and trees and shrubs in Byrsonimoideae (Anderson 1977).Climbers are plants that organize their growth strategy upon whatever rigid support, even themselves.The literature proposes two different terms to identify climbers: vines for herbaceous climbers, and lianas for woody climbers.Evolutionary studies usually consider climbers as the most derived species (Radford et al. 1974), with shrubby species as intermediate and trees as the least derived taxa (Carlquist 1991).
Few studies have attempted to compare the adaptive features in plant structure with variation in chromosome structure.Lombello & Forni-Martins (1998a) observed a relationship between the climbing habit and some trends in chromosome morphology in Sapindaceae.This pattern of karyotype derivation consisted of a decrease in chromosome number and an increase in chromosome size.
Based on chromosome numbers reported for Malpighiaceae species, this paper aims to investigate the existence of a relationship between the climbing habit derivation, as well as the fruit type derivation and the inter-specific and inter-generic basic chromosome number variation in the family.

Material and methods
Chromosome numbers of Malpighiaceae species are listed in Bolkhovskikh et al. (1969), Moore (1973;1974;1977), Goldblatt (1981;1984;1985;1988), Goldblatt & Johnson (1990;1991;1994;1997) and Lombello (2000) (Tab.1).Out of the 128 species with chromosome number report, 69 species (54%)  (1928), nomenclature on Morton (1968) and Anderson (1977Anderson ( , 1993)), with studied species habits (C-climbers, N-nonclimbers), fruit types (W-winged, N-non-winged), percentage of studied species in the genera (Sp%) and chromosome number (2n) as well as its frequency in each genera (F i,g  These chromosome data allowed us to calculate the relative frequency (F i,g ) of a given chromosome number i in a genus g with T studied species, through the following formula: Where S i = number of species with a same habit with i chromosomes in a genus g.
The F i,g value were used to calculate the weighted relative frequency (Rf i ) of each chromosome number in the two sub-families of Malpighiaceae, as well as for climbers or nonclimbers species, according to the following formula: Where G= number of calculated frequency for a given chromosome number i, and L= total frequency data for all species with the same habit.These formulas were used for the first time in relative chromosome number frequency determination by Lombello & Forni-Martins (1998a) in their study on Sapindaceae.
Taxonomic and morphological literature (Niedenzu 1928;Anderson 1977;1979;1983;1985;1993;Gates 1982;Mamede 1990;Vogel 1990) were used to characterize the habit and fruit types of the Malpighiaceae genera studied.Some of them, e.g.Banisteriopsis, Gaudichaudia and Peixotoa, present more than one habit.For these genera, the relative chromosome number frequency was calculated for each habit separately.For genera such as Camarea, that present varied habits, the chromosome counts considered in Tab. 1 were those available, even when they only accounted for species with the same habit.
In order to analyze the relationship between the fruit type derivation and chromosome number variation we distinguished between winged and non-winged fruits.The methodology used for the relative chromosome number frequency distribution was the same as that used in habit derivation analysis.

Results and discussion
Fig. 1 shows the distribution of the weighted relative frequency of each chromosome number for the two sub-families of Malpighiaceae.This figure shows a distribution strongly based on x=5 and x=10 for Malpighioideae species, and on x=6 for Byrsonimoideae.The diploid number 2n=20 represents 42,2% of the chromosome counts of Malpighioideae.For Byrsonimoideae the most significant diploid number is 2n=24 with 39,1% of the counts.Fig. 2 shows the distribution of chromosome numbers versus climbing habit for the two subfamilies.Again the diploid number 2n=20 is by far the most important for this habit, with 50,25% of the counts.The non-climber species also presented a peak for 2n=20, with 16,75% of their counts.These species presented peaks in 2n=12 and 2n=24, the latter representing the most important diploid number of the group, with 20,74% of the counts.
Adding the frequency of numbers based on x=6, and comparing the result with the total amount of numbers based on x=5 and 10, showed that x=6 derived numbers account for 62,82% of all non-climbers counts.The diploid number 2n=18 reported for Acridocarpus by Mangenot & Mangenot (1962) and Can & McPherson (1986), which could be considered as a triploid based on x=6, is here understood as an aneuploid derivation of 2n=20, based on its winged fruits, which Anderson (1993) considers derived from Mascagnia.
The results observed by Lombello & Forni-Martins (1998a) for Sapindaceae showed lower chromosome numbers in more derived species, i.e., those with climbing habits.The present paper reveals peaks of chromosome numbers in both habit types (Fig. 2).The relationship between the habit derivation and the basic number variation is evident, although non-climber species also have a significant peak at 2n=20.This requires further studies and considerations.Climbers only presented one reported chromosome number based on x=6, considering that 2n=42, observed for Hiptage benghalensis (Roy & Mishra 1962), is probably a derivation of a tetraploid 2n=4x=40.This only cited climber species belongs to the genus Stigmaphyllon, in which only 10% of the chromosome number reports are 2n=24 (Tab.1).
We also observed that the genera Banisteriopsis and Peixotoa present climbers and non-climbers species, all based on x=5 and 10.Some shrubby or arboreal species of these genera contributed to the 2n=20 peak (Fig. 2).Since some genera present habit diversity, we must consider the possibility that habit diversity occurs in some genera here classified as nonclimbing, like Heteropterys.This occurrence would certainly reduce the significance of the 2n=20 peak in the non-climbers curve.
The distribution of chromosome number frequency related to fruit type variation is similar to that observed above (Fig. 3).However, the winged fruit character is not necessarily linked to climbing habit, as seen for Dicella (Tab.1).Whilst, in genera with non-winged fruits, we almost always observe species based on x=6.Exceptions are Macvaughia and Barnebya, that show climbers based on x=10.
Given the strong relationship between basic chromosome numbers x=5 and 10 and winged fruits, and between x=6 with non-winged fruits (Fig. 2), we can interpret that the presence of winged fruits as a derived character in Malpighiaceae is a general trend.Since Malpighiaceae is a monophyletic group (Vogel 1990;Judd et al. 1999), we can hypothesize that the derivation of basic chromosome numbers x=5 and x=10 is closely related to the evolution of climbing habit and winged fruits.
Even with few chromosome data for the family (no more than 10% of the species have chromosome number reports), the importance of polyploidy is quite clear for Malpighiaceae, since about 75% of the chromosome number reports are polyploid (Tab.1).
Although Tab. 1 presents some exceptions, the similar chromosome number distribution observed for the two morphological characters analyzed (Figs. 2 and 3) reveals that Malpighiaceae is a model of relationship between evolution of morphological features and chromosome number variation.The sub-family Malpighioideae is characterized by species with climbing habit, winged fruits and basic chromosome numbers based on x=5 and 10.Byrsonimoideae shows predominance of species with arboreal or shrubby habits, non-winged fruit and basic chromosome number x=6.These latter morphological features are related to the least derived basic number in the family.

Figure 1 .
Figure 1.Distribution of weighted relative frequency (Rf i,g ) of each chromosome number for the two sub-families of Malpighiaceae.!Malpighioideae; " Byrsonimoideae.

Figure 2 .
Figure 2. Distribution of weighted relative frequency (Rf i,g ) of each chromosome number for climber and non-climber genera of Malpighiaceae.!Climbers; " Non-climbers.

Figure 3 .
Figure 3. Distribution of weighted relative frequency (Rf i,g ) of each chromosome number for winged and non-winged fruit genera of Malpighiaceae.!Winged; " Non-Winged.

Table 1 .
Sub-families, tribes and genera of Malpighiaceae with chromosome counts.Systematic based on Niedenzu ).