Sex ratio in Sand Martins (Riparia riparia) in relation to hatching date and population level

Introduction and material

In a review of sex ratio variation in birds, Clutton-Brock (1986) found little evidence of deviations from a sex ratio 50:50 at hatching. On the other hand he concluded that differential, size-related nest mortality between sexes, as well as female manipulation of the clutch was likely to create biased sex ratios among fledged young. With increasing distance to hatching date other differential effects (e.g. differences in migration behaviour) will add new bias, and an original, basic sex ratio tends to get erased and totally inaccessible. In field studies based on birds ringed as nestlings the pattern of settling distances (Berndt & Sternberg 1965, 1966, 1969, Rheinwald 1975) in first-time breeders (and of birds resettling in new areas) must be considered; if one sex disperses more than the other, this could be the main cause of biased sex ratios where these are calculated from recoveries of ringed nestlings or newly fledged juveniles.

The breeding strategy of multi-clutch Sand Martins in Northern Europe is based on successive polyandry (Cowley 1983), hence there is a need for a surplus of males, and such a surplus could be expected to occur already at hatching. Any assessment of this basic or original sex ratio must include an evaluation of settling distances and of possible differential mortality in breeding birds, however. A material, allowing such calculation has been collected in SW Scania, Sweden, since 1962. It is based on more than 69.000 Sand Martins, ringed in colonies and at roosts, and contains some 10.350 individual retraps between seasons up till and including 1999. The study area is shown in Fig. 1.

Fig. 1 Map showing the study area in SW Scania where Sand Martins have been ringed and subsequently controlled. From Persson 1987.

The dispersal of Scanian breeding-birds and first-time breeders between seasons has been treated by Persson (1978, 1987); see Fig. 2. Note that the South Scanian investigation area takes up no more than a quadrant of a full circle, with the main colonies situated close to the angular point and fifty to hundred kilometers of water extending over the three other quadrants! This topographical feature will highly facilitate the search for dismigrating birds; the effective angle of dispersal from the main colony of study, Maglarp, is no more than 100º. (The analysis of Sand Martin dispersion will be brought up to date and reformulated with the total material as point of departure.)

Fig. 2 Settling distances of first-time breeders (2k) and adults (3k+) in SW Scanian Sand Martins, shown by the cumulative total of recoveries, expressed on a percentage basis. Only controls from consecutive seasons considered. From Persson 1978.

So far the control material up till and including 1986 is available for analysis (the rest will be made available soon). It contains 1117 Sand Martins ringed as 1c birds in colonies and retrapped in the study area as breeding birds of unequivocal sex in later years. Birds were sexed from the brood patch; in Scania - with caution and experience - this method will give perfect separation as soon as females have started brooding (20 May - 15 June). And even at August roosts females moulting their brood-patches will still be recognizable as such for 1 - 2 weeks.

In Tables I - III a preliminary analysis of this material is presented. The partitioning used has an obvious heuristic background; over a sequence of years biased sex ratios and fluctuations in this bias were noted, and as years went by it became more and more evident that population level and the general progress of breeding in a single year were at the root of the problem. So, the division of the material into 2 x 3 parts is only a first, crude attempt to get to the causes behind observed fluctuations. This partitioning may - to say the least - be unsatisfactory, at the same time it may lead a few steps in the direction of a more rewarding division, perhaps with more or other parameters involved.

Table I. Sex ratio in Sand Martins ringed as 1y between 23.6 and 20.7 in periods of increasing (1963-68 and 1974-79) and decreasing (1969 - 72 and 1980-86) population and retrapped in the investigation area between 1964 and 1987. Corrected chi-square between the two cells = 5.36, p=0.02.
	PERIOD OF INCREASE "First brood", ringed 23.6-20.7	sex ratio	G-test	PERIOD OF DECREASE "First brood", ringed 23.6-20.7	sex ratio	G-test
Controlled as:	MALE: 247 FEMALE: 194	1.27	G=6.13; 1 df p<0.02	MALE: 144 FEMALE: 71	2.02	G=25.28; 1 df p<0.001

Table II. Sex ratio in Sand Martins ringed as 1y between 21.7 and 30.7 in periods of increasing (1963-68 and 1974-79) and decreasing (1969 - 72 and 1980-86) population and retrapped in the investigation area between 1964 and 1987. Corrected chi-square between the two cells = 7.20, p<0.01.
	PERIOD OF INCREASE "In between" broods, ringed 21.7-30.7	sex ratio	G-test	PERIOD OF DECREASE "In-between" broods, ringed 21.7-30.7	sex ratio	G-test
Controlled as:	MALE: 112 FEMALE: 44	2.55	G=29.73; 1 df p<0.001	MALE: 51 FEMALE: 43	1.19	n.s. from 50:50

Table III. Sex ratio in Sand Martins ringed as 1y after 30.7 in periods of increasing (1963-68 and 1974-79) and decreasing (1969 - 72 and 1980-86) population and retrapped in the investigation area between 1964 and 1987. Corrected chi-square between the two cells = 0.39.
	PERIOD OF INCREASE "Second brood", ringed after 30.7	sex ratio	G-test	PERIOD OF DECREASE "Second brood", ringed after 30.7	sex ratio	G-test
Controlled as:	MALE: 80 FEMALE: 56	1.43	n.s. from 50:50	MALE: 40 FEMALE: 35	1.14	n.s. from 50:50

The overall sex ratio in all years was 674 : 443 = 1.52; G = 48.12 (1df), p<0.001.

Discussion

The breeding of Sand Martins in the highly urbanized country of SW Scania is full of dangers and interruptions; a smooth sequence of broods without any delay is a rare occurrence. The three most common disturbances are:

Cold June with prevailing winds from NW; females are forced to abandon eggs, broods die in the burrows.
Spontaneous or man-induced sandslides throughout the season.
Predation, mainly by Badger Meles meles. In addition badgers may trigger sandslides, while prolonged predation may lead to desertion of vulnerable parts of the colony.

If there is an initial breeding synchronization (Emlen & Demong 1975) in a colony, it will of course be disrupted by such events; after some time of recovery most birds excluded from breeding will start anew at the old site, while others will emigrate and make a new breeding attempt at another colony. But apart from occurring disturbances, the original breeding synchronization is seldom more than partial. Sand Martin colonies are segregated (Sieber 1980, Persson 1987), and probably not primarily according to age, but rather according to fitness, which in turn is to a large extent, but not exclusively, age-related. A mixture of older birds and early arriving, highly fit first-time breeders are the first to breed (in a section that may be termed the "main" colony), but also the first to be struck by bad June weather and early predation. In the worst summers the early breeding is practically improductive and the first juveniles are not fledged until 10 - 15 July. These in turn come from late arriving birds, mainly first-time breeders which do not lay until early June or mid-June. The breeding sites of such late breeders may be termed marginal; they are situated at the outskirts of the colony or in sections more vulnerable to predation than the "main" colony.

My preliminary hypothesis is that the sex ratio in the offspring is determined by the fitness (and therefore to a high degree by the age) of the parents or of the female alone. Looking at the left side of tables I - III, the early breeding in years of increase - with limited negative influences from weather - has produced c. 13 males in 10 females (almost significant), which is probably a bit below the average sex ratio in any colony. The latecomers in contrast produced 25 males in 10 females, this ratio highly significant. The August breeding should again rely on a cohort of birds with high fitness; the calculated sex ratio 14:10 (not significant) is probably close to the overall average in years of increase.

Now, turning to the right side of tables I - III, the early broods often get spoiled due to unfavourable weather in years of decline. June laying increases the chances to get offspring in such years, so the first fledged young will come from parents with relatively low fitness: 19 young in 10 females, the ratio highly significant. In addition the difference between the ratios of early broods from periods of increase and periods of decrease is almost significant, and the difference between the ratios of "in-between" broods is significant. For the rest of the period in years of decrease the breeding business is carried by birds with high fitness, the resulting ratios do not significantly differ from 1:1. It should be added, that adult female mortality between seasons will double or more in winters with catastrophic mortality (Persson 1987), as a consequence the population will contain little more than first-time and second-time female breeders at the low points. As the season advances only the most fit birds will remain, while disheartened and weakened first-time breeders should be in majority among the adults migrating already from Mid-July. (Cf. Mead & Harrison 1979a, b) Also see: The shift of SW Scanian Sand Martins (Riparia riparia) from colonies to roosts in late summer

Why then do fit Sand Martins raise relatively more females than do less fit Sand Martins - if fitness is really the underlying factor? There is no size dimorphism whatsoever in the species, juvenile males and females have equal average weights and equal average wing-lengths. In consequence there should be no physical constraints likely to favour the production of either sex during the breeding season. More likely, the biased sex ratio constitutes a calculated compliance with the system of successive polyandry, in addition it may to some degree anticipate or govern future population development. How the gradual shift from producing a substantial surplus of males to 50:50 ratios (or even less) is accomplished is open to question; the whole process may be linked to physiological maturation in the breeding bird, but egg weight and its connection with female fitness is as likely to be the explanation.

This presentation is tentative and has no other aim; it merely wants to advertise: here is one large material with known rates of dismigration, where sex ratios significantly differ from 1:1. My own most urgent question or request goes: Is there any other Sand Martin material leading to similar conclusions, or are there materials concerning other species with systems of successive polygyny or polyandry, where similar ratios occur? Following that, and equally important: is there any material giving a hint to whether age (physiology) is the governing factor, or whether some category of birds is to some degree manipulating the sex ratio of its offspring? The particular case of the Sand Martin at first sight seems to be a setback to the suggestion by MacArthur (1965), that it is most favourable for a given individual to produce offspring of the minority sex, the sex in short supply. A Sand Martin female certainly transmits more genes to one following generation than her two or three male partners - but her life expectancy is lower, particularly in years of catastrophic winter mortality (Persson 1987), so in the worst case her contribution may be equal to or lower than that of the males. This could mean that there is a critical option for the individual, and that the final choice involves a complex "calculus" taking into account both population level and population development.

It seems imprudent to press the whole matter more at this stage; the preliminary ratios show beyond all doubt, that here is a problem deserving further attention. One weak point is that the overall material contains some material from the northern parts of the investigation area, where a proportion of the first-time breeding females will escape control (as will some males). Until corrected the sex ratios presented will all be a little on the high side, but this defect is not crucial and can be overcome with material from 1987 - 99. At any rate there are no major losses of females to "overseas" breeding colonies in Denmark and Germany; there are only one or two recoveries of 2y females in that direction. Immigration from or emigration to counties north of Scania (half a dozen recoveries) seem to have been connected with large sandslides in most cases, and the birds were ringed as adult, breeding birds.

LITERATURE

Berndt, R. & H. Sternberg (1965): Schematische Darstellung der Ansiedlungs-Formen bei weiblichen Trauerschnäppern (Ficedula hypoleuca). J. Orn. 106: 285 - 294. - Berndt, R. & H. Sternberg (1966): Der Brutort der einjährigen weiblichen Trauerschnäpper (Ficedula hypoleuca) in seiner Lage zum Geburtsort. J. Orn. 107: 292 - 309. - Berndt, R. & H. Sternberg (1969): Alters- und Geschlechtsunterschiede in der Dispersion des Trauerschnäppers (Ficedula hypoleuca). J. Orn. 110: 22 - 26. - Blank, J. L. & V. Nolan (1983): Offspring sex ratio in redwinged blackbirds is dependent on maternal age. Proc. Natl. Acad. Sci. 80: 6141 - 6145. - Clutton-Brock, T. H. (1986): Sex ratio variations in birds. Ibis 128: 317 - 329. - Cowley, E. (1983): Multi-brooding and mate infidelity in the Sand Martin. Bird Study 30: 1 - 7. - Emlen, S. T. & N. J. Demong (1975): Adaptive significance of Synchronized Breeding in a Colonial Bird: A New Hypothesis. Science 188: 1029 - 1031. - MacArthur, R. H. (1965): Ecological consequences of natural selection. In T. H. Waterman and H. J. Morowitz (eds.): Theoretical and Mathematical Biology, pp. 388 - 97. New York. - Mead, C. J. & J. D. Harrison (1979a): Sand Martin movements within Britain and Ireland. Bird Study 26: 73 - 86. - Mead, C. J. & J. D. Harrison (1979b): Overseas movements of British and Irish Sand Martins. Bird Study 26: 87 - 98. - Persson, C. (1978): Backsvalans dispersion i sydvästra Skåne. Anser, suppl. 3: 199-212. Lund 1978. - Persson, C. (1987): Age structure, sex ratios and survival rates in a south Swedish Sand martin (Riparia riparia) population, 1964 to 1984. J. Zool., Lond. (B)(1987), 1, 639 - 670. - Rheinwald, G. (1975): The pattern of settling distances in a population of House Martins Delichon urbica. Ardea 63: 136 - 145. - Sieber, O. (1980): Kausale und funktionale Aspekte der Verteilung von Uferschwalbenbruten (Riparia riparia L.). Z. Tierpsychol. 52: 19 - 36.

This article was written by Christer Persson, first English version published on the web: 10.8.00.