DEPRESSION BY INBREEDING AFTER FOUR SUCESSIVE SELF-POLLINATION SQUASH GENERATIONS

The strategy for breeding F 1 hybrid squash is to develop parental lines through self-pollination. However, it increases plant mean homozygosis, which is not the natural genetic state of a cross-pollinated species, and can cause “inbreeding depression”. The objective of this work was to evaluate this depression with sucessive generations of self-pollination (without selection) in Cucurbita moschata, cv. Piramoita. Populations were obtained from lines with one to four generations of self-pollination (obtained by the SSD method), from the original cv. Piramoita (population S 0 ). Randomized blocks were used with five treatments (different generations of self-pollination – S 0 to S 4 ), six replicates and five plants per plot. Regression analysis was made by the Wright inbreeding coefficient (F) to measure the homozygosis level effect on vigor loss. There was a linear reduction of mean weight and fruit length, seed production (number and weight) per fruit with the increase of the homozygosis level; however inbreeding did not affect seed quality (weight of 100 seeds and germination).


INTRODUCTION
The strategy for breeding F 1 hybrid squash is to develop parental lines through self-pollination (Whitaker & Robinson, 1986;Della Vecchia et al., 1993;Robinson & Decker-Walters, 1999;Cardoso, 2001;Maluf, 2001).However, self-pollination increases plant mean homozygosis, which is not the natural genetic state of a crosspollinated species, and it can cause "inbreeding depression".This is clearly observed in hybrid corn production, when plant size and vigor of self-pollinated inbred lines are drastically reduced as compared to outbreeding cultivars from which they originated (Allard, 1971).
Although most cross-pollinated species have higher or lower levels of inbreeding depression as a con-sequence of inbreeding, there are some in which self-pollination can happen in a continuous way with no vigor loss.Cucurbits, being cross-pollinated, are an example of a group of species in which certain lines seem to lose little vigor by inbreeding (Allard, 1971;Whitaker & Robinson, 1986;Robinson, 1999).
Despite the absense of inbreeding depression, the use of cucumber hybrids is very common, and inbred lines may be used directly as cultivars, when the objective is fruit productivity and quality (Cramer & Wehner, 1999).In some aspects, research results are contradictory regarding loss of vigor in inbred cucurbits.Self-pollination for 10 generations of Cucurbita maxima did not affect vigor or reproductive capacity (Cummings, 1928) (1922) observed that vigor loss during self-pollination did not necessarily occur.
Although several authors assume the hypothesis of reduced inbreeding depression in Cucurbita, some researchers have shown inbreeding depression for several characteristics of C. pepo and C. maxima (Borghi et al., 1973;Chekalina, 1976).Johannsson et al. (1998) reported decreased fruit yield and pollen performance by inbreeding of C. texana.Inbreeding may also affect pollen quality of C. pepo (Stephenson et al., 2001).
Cucumber, squash, melon and watermelon lines with no vigor loss have been developed.Inbreeding depression is not a limiting factor for hybrid seed production of cucurbits, in spite of being cross-pollinated.However, some lines can have decreased vigor and yield (Robinson, 1999).The present work evaluates the inbreeding depression with sucessive self-pollination for Cucurbita moschata, cv.Piramoita.
Five treatments were arranged in randomized blocks with six replicates of five plants.Treatments were: population S 0 (cultivar Piramoita) and four inbred "populations" obtained from sucessive self-pollinations (S 1, S 2, S 3 and S 4 ), with no plant selection.The single seed descent (SSD) method was used.Each population was obtained by mixing seeds of all progenies existing in each generation of self-pollination.Each progeny participated with the same number of seeds in order to obtain representative "populations" of each generation.
Seeding was made on January 3, 2002 and the transplant on January 23, 2002, using a 2.0 x 1.0 m spac-ing.Two harvests were carried out when the fruits were cream-colored, defined as morphological ripening, on April 10 and 17.Before seed extraction, the fruits were stored during seven days to complete seed ripening.
Fruit mean weight, total, neck and bulbous seed cavity length (obtaining the ratio neck length/bulbous length), seed yield per fruit (number and weight), weight of 100 seeds and % seed germination were determined.After analysis of variance, the means were compared by the Tukey test (P = 0.05).For the characteristics with statistical difference among the treatments (inbred populations), regression analysis was carried out using the Wright's inbreeding coefficient (F) to measure the homozygosis level effect on the assessed characteristics (Rubino & Wehner, 1986).

RESULTS AND DISCUSSION
Fruits presented much lower mean weight for all inbred populations than in the original population S 0 as a consequence of reduced fruit length (Table 1).Such reduction was higher for neck length and lower for bulbous length, where seeds are located.This can be noticed by the reduction in the ratio between neck and bulbous length (Table 1).Chekalina (1976) had already reported a reduction in fruit weight of different cultivars of C. maxima and C. pepo, as a result of inbreeding depression after three generations of self-pollination.
Inbred populations also presented reduced seed yield (Table 2).However, seed quality was not affected, and seeds had the same 100-seed weight and percent germination as the S 0 population.Inbreeding was harmful to seed yield but not to quality.Schuster et al. (1974), cited by Robinson (1999), had also observed that the most significant inbreeding depression effect on C. pepo was reduction in seed yield.
Vigor loss was observed in the first generation of self-pollination, contrarily to Chekalina (1976) who reported reduced vigor only after the second or third generations of inbreeding.Considering an increment of 50%   in homozygotes in each generation of self-pollination, higher loss in the first generation, decreasing in subsequent generations, is expected.This trend was observed in all characteristics with vigor loss, since the results followed a linear regression model, with values decreasing with increasing homozygosis, evaluated through Wright's coefficient (Table 3).Little vigor loss by inbreeding in cucurbits is attributed to small population maintenance because of their need for space.Consequently, this results in inbreeding in spite of the fact that floral biology favors cross-pollination (Allard, 1971).However, the studied population was originated from an interspecific cross (C.pepo x C. moschata), involving three cultivars (Yankee Hybrid, Butternut and Menina Brasileira), without self-pollination in the end of the program (Costa, 1974), a fact that probably resulted in great genetic variability and high quantity of heterozygous loci.The incorporated bush character favors an increase of plant number per area, reducing the chance of natural self-pollinations during seed maintenance and multiplication phases.This may be the reason for high depression in this population.
By the single seed descent method, lines obtained will hardly be more productive than the original population.However, mean of the populations with different inbreeding levels were assessed, because lines obtained after a certain number of generations of self-pollination may present different results in relation to vigor loss (Maluf, 2001).This experiment evaluated the progeny mixtures, which presented reduced mean weight and fruit size, and lower seed yield.Nevertheless, there may exist progenies in which there is no loss as compared to the original S 0 population.Moreover, low-vigour, inbred progenies may generate populations or vigorous hybrids after heterozygosity recovery.Seed yield, however, can be damaged by lower seed production in lines, affecting their final cost.Vigor loss in Cucurbita moschata, cv.Piramoita, with successive generations of self-pollination, does not prevent its use for obtaining lines in breeding programs.

Table 1 -
Fruit mean weight, total, neck and bulbous seed cavity (BSC) length, and relation between neck and bulbous length in populations with different inbreeding levels.SãoManuel-SP, 2002.

Table 2 -
Weight of 100 seeds, seed yield and quality of populations with different inbreeding levels.São Manuel-SP, 2002.1 Means followed by the same letter do not differ by the Tukey test (5%) 2 Coefficient of variance