Extreme temperatures compromise male and female fertility in a large desert bird

Abstract

Temperature has a crucial influence on the places where species can survive and reproduce. Past research has primarily focused on survival, making it unclear if temperature fluctuations constrain reproductive success, and if so whether populations harbour the potential to respond to climatic shifts. Here, using two decades of data from a large experimental breeding programme of the iconic ostrich (Struthio camelus) in South Africa, we show that the number of eggs females laid and the number of sperm males produced were highly sensitive to natural temperature extremes (ranging from -5 °C to 45 °C). This resulted in reductions in reproductive success of up to 44% with 5 °C deviations from their thermal optimum. In contrast, gamete quality was largely unaffected by temperature. Extreme temperatures also did not expose trade-offs between gametic traits. Instead, some females appeared to invest more in reproducing at high temperatures, which may facilitate responses to climate change. These results show that the robustness of fertility to temperature fluctuations, and not just temperature increases, is a critical aspect of species persistence in regions predicted to undergo the greatest change in climate volatility.

Fig. 1. Ostriches (Struthio camelus) cope with large thermal fluctuations in their native habitat, reproducing successfully across Africa from the Western Cape to the deserts of Southern and Northern Africa.

a Courtship by a male ostrich (right) towards a female (left) in one of the enclosures (n = 197) at the study site used to keep a single breeding pair (photo: CKC). b Data structure of fertility traits obtained from 1998 to 2018 at the study site of Oudtshoorn Research Farm in the arid Klein Karoo region of South Africa. Sperm viability data was not available for all of the solitary males where measures of sperm numbers were obtained (sperm viability: n_ind = 18, n_years = 7, ${\overline{x}}_{\mathrm{years}\mathrm{per}\mathrm{ind}}=2.7$). See also Supplementary Table 1 for detailed overview of sample sizes. c Geographic range (green) of the ostrich with the study site marked by an asterisk. d Monthly temperature range was calculated by estimating the range of temperatures of each month and then calculating the mean of this across all months.

Fig. 2. Temperature extremes compromise male (n = 22) and female (n = 652) fertility.

Female egg laying rate (a) and number of sperm ejaculated by males (b) were both highly sensitive to increases and decreases in temperature. Female (c: egg mass) and male (d: sperm viability) gamete quality were generally more resistant to temperature change. Hatching success (e), which is influenced by the egg mass^,, sperm numbers and sperm viability, was also less affected by temperature change. The number of offspring (f) is a product of hatching success and rates of egg laying and was influenced by changes in temperature that occured during egg laying. Ostrich females can only lay an egg every other day and we therefore used number of eggs or chicks per number of two-day intervals (eggs/2 days or chicks/2 days) (see the subsection “Time lag effects of temperature on gametes” in “Methods” section). The range of temperatures that sperm traits were measured at differed from the other traits, because it was not possible to collect sperm across all years (Supplementary Table 1). Fitted lines and 95% credible intervals (shaded area) from the primary set of models are shown for traits significantly affected by temperature (Supplementary Tables 2–7). For binomial models the fitted lines span the modelled binned temperature classes making them robust to outliers. Points are averages with standard errors binned according to the temperature variable. Point size illustrates relative number of observations. Source data are provided as a Source Data file.

Fig. 3. Correlated changes in the number and quality of gametes as temperatures increased and decreased.

The number of eggs and sperm females (n = 652) and males (n = 18) produced was not traded-off against egg mass and sperm viability as temperatures changed (see also Supplementary Table 18). This was consistent within and among individuals. Changes in egg-laying rates were positively correlated to egg mass as temperatures increased both within and among females (credible interval (CI) of phenotypic correlation excluded zero). Source data are provided as a Source Data file.