Resource levels, reproduction and resistance to haematozoan infections
Ju « ki « rgen Wiehn* and Erkki Korpima
Section of Ecology, Department of Biology, University of T urku, FIN-20014 T urku, Finland Compromise of immune function during reproduction may form a link between parental e¡ort and the cost of reproduction, but the role of environmental variation in structuring intra-individual life-history trade-o¡s has been poorly investigated. We manipulated the need for parental e¡ort in Eurasian kestrels, Falco tinnunculus, by food-supplementing broods for three nestling periods, during which the natural main food supply (voles) varied, and found that parental parasitaemia was inversely related to yearly vole densities. The level of parasitaemia in females was, however, reduced by food supplements. No e¡ect on males was expected, as earlier work has shown that only females responded to the supplements by changing their behaviour. We show directly that the likelihood of female parasitaemia was diminished by spending less time in £ight-hunting, which was related to reproduction during a good vole year, to our supplementary feeding, and to being mated to a male with high parental e¡ort. Our results represent a novel direct bene¢t for females in resource-providing species, linked to female as well as o¡spring, wellbeing, and they provide insight into why the appearance of reproductive costs may be linked to gender or environmental conditions. Keywords: cost of reproduction; direct bene¢ts; Falco tinnunculus ; haematozoa; life-history trade-o¡s; parasitic infections tozoa have experimentally been shown to cause histopathological changes in hosts (Molyneux et al. 1983; Atkinson et al. 1988), and, in particular, have potentially been shown to reduce the ¢tness of kestrels (Korpima « ki et al. 1995). In a previous study we showed that females of food-supplemented nests decreased their hunting e¡ort and provisioning of their nests, whereas males continued with unchanged levels of hunting and provisioning. Similarly, females hunted little if their mates were active in provisioning the brood (Wiehn & Korpima « ki 1997). Therefore, in addition, to study the e¡ects of resource manipulation, we directly assess the relationship between hunting e¡ort and parasitism in this study, and evaluate the consequences of reduced male provisioning for female susceptibility to parasitaemia.
2. METHODS We studied nestbox-breeding kestrels in an agricultural area of western Finland (62859'^63810' N, 22850'^23820' E), where migrant kestrels arrive from late March to early May (Palokangas et al. 1992) and exhibit low breeding site tenacity (yearly mean turnover of breeding males 75%, and of females 92%, during 1983^1992 (Korpima « ki 1988; J. Wiehn & E. Korpima « ki, unpublished data)). The kestrel is a widespread open-country raptor, which shows a clear sexual division of labour; males provide most of the food between the pair-formation and mid-nestling phase, and females incubate and brood the young exclusively and start to hunt along with their mates only when nestlings are around two weeks old. Their main hunting modes are wind-hovering, soaring and perching (Village 1990).
& 1998 The Royal Society
That reproduction can entail a cost has become clear (Lessells 1991). Recently, reduced immunocompetence as a consequence of increased reproductive e¡ort has been seen as an important pathway for the cost of reproduction (see Sheldon & Verhulst (1996) for a review). However, the role of environmental variability in the context of intra-individual trade-o¡s has been poorly studied. The connection between high current reproductive e¡ort and low residual reproductive value assumes the ratio of investment between reproduction and functions that are likely to increase future survival or fecundity to be always of same magnitude (Tuomi et al. 1983). This might not always be the case, as environments usually vary in quality over time or space, and hence individuals breeding in good-quality environments might not face as large allocation trade-o¡s as those breeding in poorquality environments. T o study the role environmental variability can have in structuring intra-individual life-history trade-o¡s, we conducted a supplementary feeding experiment on Eurasian kestrels, Falco tinnunculus, (hereafter kestrel) breeding in Finland, in a highly variable environment with respect to the yearly supply of their main food, voles (Korpima « ki 1985). Natural-sized kestrel broods were food-supplemented for the whole nestling period over three years in which the abundance of voles varied, and the prevalence of parental intracellular Haemoproteus and extracellular Trypanosoma haematozoa was determined. These haema*
Author for correspondence (jyrwie@utu.®).
Proc. R. Soc. Lond. B (1998) 265, 1197^1201 Received 19 March 1998 Accepted 3 April 1998
we supplemented nests every other day for the whole duration of the nestling period by providing them with dead rooster chicks (ca. transferred to a clean glass slide. respectively (for detailed indices. and chose the model with best ¢t using backward (stepwise) model selection.Wiehn and E. p 0.) For determining haemoproteids. (H. 2. Bennett by screening 100 ¢elds under oil. these years can be regarded as `low'. 1. ¢rst with one (young kestrels 41 week) and later with two (young 41 week) per kestrel nestling.. d. p 0. Canada. blood was collected in 60 ml haematocrit tubes. gender Â treatment Â year (LR.f. 1 in Wiehn & Korpima « ki (1997)). which divide and circulate in the bloodstream. which are. p 0.m. wind-hovering.44. thus covering approximately half of the daily requirement of nestlings (Masman et al. d.27. Brie£y. B (1998)
(b) Blood sampling
As the infections were treated as binary variables (noninfected or infected). Pen. 12 5. soaring and gliding) and preydelivery rates to nests (prey items per hour). see ¢g. logistic regression was used to estimate infection probabilities of kestrels. Experiments were conducted for three years. we calculated individual hunting e¡ort (proportion of time spent in energetically expensive (Masman & Klaassen 1987) £ight-hunting.0217). Immediately thereafter. In 1993. p 0. directional £ight. (a. Trypanosoma. Korpima « ki. with indicator (dummy) coding. a drop of blood was collected in a microcapillary tube. Nests with matched clutch sizes and hatching dates were identi¢ed as controls. £agellated trypanosomes.0005). air-dried.0230).f.
(a) Feeding experiments
proteus spp.e. p 0. avium complex) and intracellular HaemoProc. d. d. (b. the tubes were centrifuged for 5 min at 5000 r. 2.
(c) Statistical analyses
J. i. treatment (LR.
Male and female parents were captured with a swing-door trap attached to the nestbox when the nestlings were about two weeks old. treatment Â year (LR. Means ( Æ binomial s. 1992 and 1995. 30 g). on average.55.f. gender Â treatment (LR.0488). brachiatus) infections.03.p. and the repeatability between screeners was good: all the individuals were determined as either infected or non-infected by both the screeners. 18% and 30% larger in years of high rather than of low vole abundance (J. were also quanti¢ed by means of a haematocrit centrifuge method (Woo 1970). tinnunculi and H. p 0. d) among male (a. 12 12. 1. F. d. We report only two-tailed signi¢cant terms. d) parents rearing young in years of low (1993). For this method.
Nests were food-supplemented as described earlier (Wiehn & Korpima « ki 1997). 12 4. 1989). Soc. directly re£ecting clutch and brood sizes. unpublished data).39. Signi¢cant terms: Haemoproteus. (No other infections were found. intermediate (1992) and high (1995) natural food (vole) supply in nests that served as controls (open circles) or were supplemented (¢lled circles) with rooster chicks for the period of nestling time. (T. `intermediate' or `high' with respect to relative vole abundance during breeding seasons 1993.W. The signi¢cance of the models was estimated using the likelihood ratio (LR) method. Lond.0147). and hunting e¡ort and prey-delivery rate as continuous covariates. Slides labelled only with a ring number were sent to the International Reference Centre of Avian Haematozoa. and in 1995 by both Ido Pen and J. Y ear and treatment (supplementation or control) were treated as categorical variables. d. year (LR. Two taxon-speci¢c methods were applied to determine the prevalence (% individuals infected) of extracellular Trypanosoma spp.) are shown. Wiehn & E.f. and ¢xed in absolute ethanol some hours later (Bennett 1970). c) and Trypanosoma spp. trypanosomes were screened by inspecting each tube for 5 min under a light microscope at Â 400 magni¢cation. R. 1 2 17. b) and female (c. 12 8. with the logistic regression procedure in SPPS for Windows software (Norusis 1993). Based on 6^8 h time budgets per pair during mid-nestling time (the young were aged 13^16 d).f.e. Within 12 h of sampling. 2. 1. Prevalence (% individuals infected) of Haemoproteus spp. Korpima « ki
Reproduction and haematozoan infections in kestrels
Figure 1. From 1993. based on the snap-trapping indices of voles. 12 7. using Â1000 magni¢cation. Experimental nests were also subject to behavioural observations. smeared.f.0002). the tubes were inspected by I. Numbers are sample sizes. year (LR. where they were subsequently inspected by G. This variation in food conditions is a general pattern at the study site. which were subsequently stored at ambient temperature. which di¡ered with respect to the natural abundance of voles.25. We entered the main e¡ects and their interactions into analyses. Blood samples from parents were taken by brachial venipuncture.
These results support ¢ndings that poor nutrition is often associated with impaired immunological defence and disease (Chandra & Newberne 1977).f.09. d.0 (Trypanosoma spp.25 (Haemoproteus). Soc. a similar relationship in females was modi¢ed by food supplements.0108. During years of better food conditions this trade-o¡ is apparently less important. Hard-working. and (b) Trypanosoma spp.0029) and 26% were infected with one of the parasites but not the other. or a newly acquired infection. 12 6. no signi¢cant relationship with male hunting e¡ort (LR.f.90. p 0. intermediate and high refer to the yearly natural food (vole) supply in 1993. the main providers for the family. Whereas the parasitic prevalence of males followed inversely yearly food conditions (¢gure 1a ^ b). but our results clearly suggest that the trade-o¡ between two competing processes. however. Note that the x-axis is o¡set slightly below zero of the y-axis to avoid overlapping of the x-axis and prediction lines. 1. d. why was this seen only in females? If the connection between hunting for nestlings and parasitaemia was straightforward. p 0. d.f. is most pronounced during unfavourable breeding conditions. in nature. Therefore.77). Whether the parasitaemia arose from only the energy allocation trade-o¡ or with di¡erent vector pressure accompanying it remains unclear. 1. infection in female (black dotted lines) and male (grey solid lines) parents as a function of hunting e¡ort (% of time spent in £ight-hunting). R. Parasitaemia in adults could arise during the nestling period in two di¡erent ways. 2. might cause an infection (relapse or new) through reduced immunocompetence or increased exposure to vectors.Reproduction and haematozoan infections in kestrels J. provisioning
. this relationship in females was modi¢ed by year.
the nestling-rearing period. 1. food supply and parental e¡ort
Figure 2. 1. By giving supplemental food. p 0. migrant birds might arrive at the breeding areas without detectable infections and.0173. But.001) for females.f. makes females susceptible to infection. Again.49.Wiehn and E. First. 1997). hunting time should also have predicted male parasitaemia. resulting in 81% lower prevalence of Haemoproteus in the year of low food supply (1993) (LR. it seems as if hunting. 12 0. 12 2. we believe that our study provides the ¢rst direct evidence for huntinginvestment-related parasitaemia in nature.f. respectively.11). even if the hunting of males was not changed by food supplements (Wiehn & Korpima « ki 1997). show increased levels of parasitaemia during
Proc.71. reproductive e¡ort and immune response. and that energetically expensive (Demas et al. The negative relationship. p 0.31. Korpima « ki 1199
3. further showing that the magnitude of the trade-o¡ between con£icting demands.f. when females had both very low levels of infection (¢gure 1) and reduced hunting times (¢gure 2).56. and increased time in energetically expensive (Masman & Klaassen 1987) £ight-hunting predicted elevated parasitaemia in females. p 0. 1992 and 1995. or both. Deerenberg et al. we could decrease females' hunting times (Wiehn & Korpima « ki 1997). for both of the parasites. or microhabitat use in hunting (and thus the probability of encountering vectors) could have di¡ered from year to year.0095) for females. p 0. vector densities could have varied. RESULTS AND DISCUSSION
We found 27% of the birds to be infected with Haemoproteus and 28% with Trypanosoma.67. p 0.
(b) Male parental investment and direct bene¢ts for females
If hunting has such a pronounced e¡ect on female susceptibility. but not in males (¢gure 2). ¢gure 1c) and 84% lower Trypanosoma prevalence in the supplemented nests when years were combined (LR. the parasitaemia is perhaps in£uenced by cumulative investment e¡ects during the entire breeding cycle. in most cases where females were infected. especially given the males' main role as provider ? Within pairs. Lond. p 0. 1. between yearly food supply and parasitaemia in males also indicates that the same amount of investment in hunting over the years (Wiehn & Korpima « ki 1997) leads to decreased ability to invest in parasite defence with declining environmental conditions. ¢gure 1d). 1997) immune response is impaired by increased parental e¡ort (Richner et al. the infection status would di¡er.f. In our study. 1 2 9. 1995. whereas in males. males were not infected (Fisher's exact probability tests. is likely to be modi¢ed by environmental conditions. 12 5. Probability of (a) Haemoproteus spp. Lines depict the predictions of regression models. Seventy-four per cent had either both of these parasites or none of them (12 8. However. d. d. d.)). such as parental investment and immunofunction. The categories of low. (a) Hunting e¡ort Â year (LR. 1. the estimates of hunting e¡ort represent the range of observed £ight-hunting times. as a result of either a relapse of an old infection (Atkinson & van Riper 1991). (b) Hunting e¡ort Â year (LR. d. no signi¢cant relationship for male hunting e¡ort (LR. in that £ighthunting predicted parasitaemia in years other than the high food year. 12 10. p 1. which in females only begins during the nestling phase. why should they hunt. B (1998)
(a) Parasites. Within a given pair.
Bennett for screening the haemoproteids. From a female's point of view. low provisioning of males increased the likelihood of parasitaemia in females (¢gure 3). Daan. J. and P. For simplicity of presentation. 1993. in that females of supplemented nests had a lower probability of infection on a given male provisioning rate than females of control nests (¢gure 3).. 12 4. pp. 1. Lines depict the predictions of regression models. G. Ka «a « r.0455).. 1977 Nutrition. & Forrester. Soc. Anim. B 264. E. Sakari Ikola for logistic assistance in the ¢eld. Therefore.
Figure 3. 12 7. 1996 Increased daily work precipitates natural death in the kestrel. 1970 Simple techniques for making avian blood smears. New Y ork: Plenum Press. J. Lond. 228^239. Korpima « ki
Reproduction and haematozoan infections in kestrels knowledge.W. the supplementary feeding simulated increased male provisioning. Moreover. 65. Daan.g. & Dijkstra. D. Apanius. 539^544. Bennett. Proc.37. Korpima « ki et al.18.f. The work was supported ¢nancially by awards from the Emil Aaltonen Foundation (to J. the ¢rst study showing that the bene¢t may come through the well-being of the female. with unique manipulation of resources within years of naturally di¡erent food conditions. B. R. D.1200
J. Deerenberg. Our results also pinpoint a potential defect of many observational studies reporting reduced reproductive output of individuals with high parasitic levels (e.
rates of males and females were negatively correlated. F. Deerenberg. M. p 0. 1021^1029. J. Ido Pen for screening the trypanosomes. Wiehn 1997. year is omitted. In e¡ect.00. Oxford University Press. T. Wiehn and E. Loye & Zuk 1991. (ii) male prey-delivery rate (LR. 1st edn. R. Sheldon and an anonymous referee for comments on the manuscript. Wiehn et al. In Bird ^ parasite interactions (ed. 1997 Reproductive e¡ort decreases antibody responsiveness. C. C. J. Ido Pen. we have demonstrated a potential cost of being paired to an individual that makes a low investment in parental care. Greiner. Ultimately. 74. Note that the x-axis is o¡set slightly below zero of the y-axis to avoid overlapping of the x-axis and prediction lines. This relationship was further modi¢ed by food supplements. immunity. Princeton University Press. C. 1988 Pathogenicity of Haemoproteus meleagridis (Haemosporina: Haemoproteidae) in experimentally infected domestic turkeys. d. this raises mate choice considerations.f. d. 19^48. the direct bene¢ts for females choosing a male with high parenting contribution have been understood as gains for the well-being of females' o¡spring (Andersson 1994). P. 48. S. 1. Leucocytozoon.. E. N. Soc. Parasitol.
. and suggests that kestrel females are likely to gain direct bene¢ts by choosing a `good parent' (Palokangas et al. the delivery rates represent the range of observed rates. B (1998)
Andersson. Probability of (a) Haemoproteus spp. S. Zool. & Bos. Atkinson. MÖller et al. C. and (b) Trypanosoma spp. J. d.). & Newberne. 12 5. Traditionally.) and the Academy of Finland (to E. infection among females as a function of male prey-delivery rate (items brought to nest per hour). Moreno. which may then enhance her own residual reproductive value.f. the late Gordon F. without experimental veri¢cation. C. Lond. 585^586.f. This is. V . This is clearly seen in our experiment: kestrel nests that were food-supplemented £edged more young each year than control nests (Wiehn & Korpima « ki 1997) and females attending these supplemented nests also showed lower prevalence of parasites than those of control nests. 1990. 1995).0366) and the male delivery rate Â treatment interaction (LR. Jussi Ryssy and especially Pasi T olonen for help in collecting data. p 0. C.0178). J. and they aid in understanding why the appearance of trade-o¡s may be linked to gender or environmental conditions (Korpima « ki & Rita 1996). the causality in relationships between reproductive output and parasitaemia should be interpreted with caution. these results suggest a mechanism for the trade-o¡ between reproduction and survival demonstrated experimentally in the kestrel (Daan et al. R. The relationships do not change qualitatively if year (categorical variable) is added to the model.K. III 1991 Pathogenicity and epizootiology of avian haematozoa: Plasmodium. 1. Signi¢cant terms are: (i) male prey-delivery rate (LR. Chandra. Consequently. the role that environmental variation plays in structuring life-history trade-o¡s. and Haemoproteus. & van Riper. which suggests that females had to increase their own hunting to compensate for their mates' poor provisioning (Wiehn & Korpima « ki 1997). These negative associations can arise when individuals di¡er in genotypic quality or as a consequence of nutritional e¡ects creating covariance between ¢tness and parasite prevalence. 1. to our
Proc. Atkinson.0074) and treatment (LR. Ecol. J.
We thank Mikko Ha « nninen. 1994. without any direct causal connection between the two. 1994 Sexual selection. d. and infection. Zuk). C. 1996). 12 4. Erkki Rautiainen. K. M. Kai Norrdahl. p 0. p 0.61. Can.. Merila « . We have experimentally demonstrated. 1997). Bernd Riedstra. Loye & M. C. Norris.
J. 1991 The evolution of life histories. C. pp. 921^923. 1970 The hematocrit centrifuge technique for the detection of trypanosomes in blood. 11. E. Woo. 659^665. 43. Lessells. & Bennett. 32^68.. J. 1987 Energy expenditure during £ight in trained and free-living kestrels. 1996 E¡ects of brood size manipulations on o¡spring and parental survival in the European kestrel under £uctuating food conditions. Zool. & Bennett. 62. 1995 Blood para« ki. J. and selection in life-history evolution. 1983 Studies on the pathology of an avian trypanosome (T. Huhta. Talan. W. & Zuk. 49^54. 2. P. 1997 Food limitation on brood size: experimental evidence in the Eurasian kestrel. E.. J. 1990 Fitness e¡ects of parasites on passerine birds: a review. M. G. 47^55. Molyneux. R. & Sorjonen. Trends Ecol. J. J. Reg. 1993 SPSS advanced statistics user's guide. R. H. R. evolution and behaviour. Oecologia 77. B. Wiehn. Behav. P. K. I.
Proc.. Krebs & N. 1989 Energetic limitation of avian parental e¡ort: ¢eld experiments in the kestrel (Falco tinnunculus). R. Avian Biol. G. Alatalo. Masman. Ecol. & Alatalo. J. & Verhulst. 1996 Ecological immunology: costly parasite defences and trade-o¡s in evolutionary ecology. E. E. Wiehn. H. V . F. Parasitology 87. Biol. E. Daan. pp.Reproduction and haematozoan infections in kestrels J. P. 603^616. Lond. costs of reproduction. Bildstein. 1997 Mate choice and reproductive success in the American kestrel: a role for blood parasites? Ethology 103. E. 130^137. 278^285. J. P. V . P. J. E. J. C. D. P. S. sexual selection and reproductive success of European kestrels. J. 1997 Plumage characteristics as an indicator of male parental quality in the American kestrel. J. In Population biology of passerine birds: an integrated approach (ed. Christe.. Auk 104. Natn. B. 335^343. C. Phys. 317^321. & Haukioja. A. E. USA 92.. Sheldon. Cooper. A. Chefer. & Oppliger. Comp. Hakkarainen. R. Evol.. E. T. J. 435^455. Berlin: Springer. IL: SPSS. Norusis. Korpima « ki. Zool. Ecology. Lebreton & R. Dijkstra. B (1998)
.Wiehn and E. Anim. Korpima « ki. 1985 Diet of the kestrel Falco tinnunculus in the breeding season. 1997 Metabolic costs of mounting an antigen-stimulated immune response in adult and aged C57BL/6J mice. Palokangas. H. 47. 2043^2050. Sci. 1988 Factors promoting polygyny in European birds of preyöa hypothesis.. K. T sites. M. A. bou¡ardi) infection in experimentally infected canaries. Falco tinnunculus. T. J. Gosler.. Am. Chicago. P.. M. D. A. H. H. 28. & Korpima « ki. Oxford University Press. V . 23. MÖller. Can. Ecoscience 3. McCleary). F. E. Korpima « ki.. Int. (eds) 1991 Bird ^ parasite interactions. Behav. 1990 The kestrel. Wiehn. J. & Smith. Evol. Anim. Ornis Fenn. Proc. Korpima « ki. Korpima olonen.. 42. Masman. Ecoscience 2. Village. 420^426. & Rita. Davies).. G.. 7. London: Poyser. J. M. Allander.. H. 443^448. D. L. 1994 Female kestrels gain reproductive success by choosing brightly ornamented males. 47... 1992 Female choice in the kestrel under di¡erent availability of mating options. E. Phys. E. & Bult. & Nelson. Oxford: Blackwell Scienti¢c. Acad. Korpima « ki. 25^34. S. 1st edn. & Klaassen. Korpima « ki 1201
Demas. Richner. 1192^1194. 264^273. T olonen. Ecology 78. Korpima « ki. R. J. In Behavioural ecology: an evolutionary approach (ed. 1993 Blood parasites and reproductive success of T engmalm's owls: detrimental e¡ects on females but not on males? Funct. Tuomi.. & Dufva. K. D. A. Soc. Hakala. E. R1631^ R1637. & Korpima « ki. M. Palokangas. E. 1995 Paternal investment a¡ects prevalence of malaria. 1983 Alternative concepts of reproductive e¡ort. 304^317. Hakkarainen. Loye. Blondel. 269^280. Am.