THE DIFFERENTIAL-ALLOCATION HYPOTHESIS: DOES THE EVIDENCE SUPPORT IT?

Abstract

In a recent paper, de Lope and M0ller (1993) investigated the differential-allocation hypothesis in a Spanish population of barn swallows Hirundo rustica near Badajoz in 1991, and interpreted the results as support for this hypothesis. I believe that another interpretation could explain their results. The differential-allocation hypothesis predicts that mates of attractive individuals are to contribute more than the average (of their sex) to parental investment in order both to obtain and keep their attractive mates (Burley 1986). Consequently, the attractive partner can afford to reduce its parental effort and thus gain a survival advantage. Offspring of pairs with an attractive individual gain an advantage by increased body mass when fledging and survive better than other offspring. This hypothesis has only been tested and supported once in a laboratory study by manipulating the attractiveness of male and female zebra finches Taeniopygia guttata, by changing the color of their leg bands (Burley 1988). De Lope and M0ller changed the attractiveness of male barn swallows by shortening or lengthening the outermost tail feathers and addressed the question of whether tail manipulation of already mated males affects the reproductive effort of their female partners. They assigned males randomly to one of three groups: Shortened males had a 20-mmlong piece of outermost feathers cut; males had 20-mm-long pieces of outermost feather added to their own tail feathers; control males were not manipulated. Since previous studies have shown that female barn swallows prefer males with artificially elongated tails (M0ller 1990), the authors conclude that males with tails become unattractive and elongated-tailed males become attractive. De Lope and M0ller made clear predictions for the differentialallocation hypothesis: they expected females mated to elongated-tailed males to have increased reproductive success, but also to make a greater provisioning effort; by comparison, they expected the provisioning rate of males to decrease with increasing tail length. The data on reproductive effort in females supported the differential-allocation hypothesis. Female mates of elongatedtailed males indeed bred earlier and produced more eggs and nestlings per nest and season. But contrary to the prediction (if females adjusted their reproductive effort to experimental tail length, females mated to long-tailed males should . . . produce nestlings of higher body mass; de Lope and M0ller 1993, p. 1155) nestlings of pairs with elongated-tailed males did not have higher body mass than those of the shortened and control group. Because the survival prospects of nestlings are positively related to body mass (Perrins 1965), nestlings of pairs with an attractive partner do not survive better. The authors estimated the feeding rates of each partner in randomly sampled pairs for 1 h on two different days when nestlings were 1O-15-d old. Additionally, they provided nestlings with a ligature and collected food boluses brought by the attending males. They recorded the number of prey items per bolus and measured the length of the prey items to the nearest 0.1 mm. They found a significant decrease in the feeding rate of shortened males over control males to males with elongated tails and a significantly increasing trend in female feeding rate in the same ordering, and concluded that these results support the differential-allocation hypothesis. However, this effect could be explained more simply by a compensation effect. Because an elongated tail is a handicap and reduces foraging efficiency (M0ller 1989) female partners of elongated-tailed males have to feed more frequently in order to provide enough food for the young. Thus, because males in the group were unable to feed more due to an impairment imposed on them, their females had to increase their feeding rates. Accordingly, females were forced to invest more to maintain the integrity of their genetic investment. The differential-allocation hypothesis, however, predicts that the females are willing to do more than the average of their sex because they have acquired an attractive partner (Burley 1986). It is difficult to distinguish between the compensationeffect and the attractiveness-effect predicted by the differential-allocation hypothesis because the feeding rate of one partner depends on that of the other partner. To bypass this difficulty, one could, for example, provide additional food to the young fathered by the elongated males up to a level at least matching the deficit imposed by those males. Assuming that the compensation-hypothesis is true, females should no longer increase their feeding rate. However, if the differentialallocation hypothesis were true, the same females would feed their young more frequently than females of the other groups, that is, they would feed more than necessary. There is yet another snag: Do the elongated-tailed males indeed reduce their feeding effort? The differential-allocation hypothesis focuses on the provisioning effort of the individual mated to an attractive partner. Because de Lope and M011er determined the attractiveness of the males, they should