Introduction
Egg incubation and hatching are the first major hurdles that many egg-laying taxa must overcome to achieve reproductive success (Deeming and Ferguson Reference Deeming and Ferguson1991; Du and Shine Reference Du and Shine2015). Failure of eggs to hatch is fairly common, particularly in threatened species, and has significant impacts on the stability and persistence of populations (Brekke et al. Reference Brekke, Bennett, Wang, Pettorelli and Ewen2010; Ferreira et al. Reference Ferreira, Hansen, Parrish, Pierce, Pulham and Taylor2005; Jamieson and Ryan Reference Jamieson and Ryan2000; Marshall et al. Reference Marshall, Balloux, Hemmings and Brekke2023; White et al. Reference White, Eason, Jamieson and Robertson2015). It is therefore important that rates and causes of egg failure and loss are accurately monitored.
Egg failure and loss occurs for various reasons, including predation, damage, climatic factors, pollution, and genetic issues that impact embryo survival (Assersohn et al. Reference Assersohn, Marshall, Morland, Brekke and Hemmings2021). In addition to these drivers, a noticeable proportion of eggs fail to hatch because they are abandoned by the parent(s), prior to or during incubation. Parents may abandon entire clutches, or a subset of the eggs (partial clutch abandonment), the latter being especially common after partial predation events (e.g. Ackerman et al. Reference Ackerman, Eadie, Loughman, Yarris and McLandress2003). We therefore use the term “egg abandonment” to refer to the abandonment of either a subset of eggs or the entire clutch, occurring at any time between egg laying and the end of the incubation period. If the parent abandons eggs after the expected end of the incubation period, these eggs should be classed as “failed” rather than abandoned, since they must have failed before the parent left. It should be noted that abandonment behaviour differs from egg ejection/expulsion (e.g. Lobato et al. Reference Lobato, Moreno, Merino, Sanz, Arriero and Morales2006), rejection, and nest sanitation (e.g. see Guigueno and Sealy Reference Guigueno and Sealy2012) behaviours, which all involve the physical removal of eggs rather than the cessation of parental care while the eggs remain in the nest.
Egg abandonment is commonly associated with disturbance, climatic, and/or environmental events (see below), so it is likely that abandonment is not usually related to egg quality or viability, and the majority of abandoned eggs therefore represent a random subset of those laid. Consistent with this, attempts at fostering abandoned eggs have previously proven successful (e.g. O’Connell et al. Reference O’Connell, Power, Keogh, McGuirk, Macey and Newton2015) and there is no substantiated evidence for preferential abandonment of unfertilised eggs or inviable eggs (Marshall et al. Reference Marshall, Balloux, Hemmings and Brekke2023). The causes of abandonment are therefore mostly assumed to be independent of other drivers of either fertilisation failure or embryo mortality, and as such, abandoned eggs are typically excluded from standard calculations of “hatching failure” along with eggs that are predated or damaged (e.g. Briskie and Mackintosh Reference Briskie and Mackintosh2004; Marshall et al. Reference Marshall, Balloux, Hemmings and Brekke2023), or, inversely, “hatchability” (Koenig Reference Koenig1982) and “hatching success” (e.g. Spottiswoode and Møller Reference Spottiswoode and Møller2004).
There are three potential problems with the exclusion of abandoned eggs from hatching failure monitoring and analysis efforts. First, methods of defining and recording egg abandonment are often inconsistent and subjective, inevitably leading to inaccurate classifications of abandonment that upwardly or downwardly bias estimates of hatching failure rates (see below). Second, the identification of egg abandonment may be impacted by nest monitoring effort, with the risk that egg abandonment is more commonly mistaken for other causes of hatching failure (or vice versa) when nest visits are sporadic. Third, in cases where abandoned eggs are assumed to be fertilised, abandonment is usually considered to be the cause of embryo mortality (but see Beissinger et al. Reference Beissinger, Cook and Arendt2005), especially since abandoned eggs often have the potential to hatch if subsequently incubated (O’Connell et al. Reference O’Connell, Power, Keogh, McGuirk, Macey and Newton2015; Schacter et al. Reference Schacter, Fettig, Peterson, Hartman, Herzog and Casazza2022). It is feasible, however, that abandonment is sometimes the consequence of embryo death, rather than the cause. In birds, for example, it is well established that parent–embryo communication occurs, particularly during the late stages of egg incubation (Brua Reference Brua and Deeming2002; Brulez et al. Reference Brulez, Pike, Reynolds, Deeming and Reynolds2015; Vince Reference Vince1966, Reference Vince and Hinde1969), and embryos actively respond to and modulate the incubation environment provided by their parents during development (Brua Reference Brua and Deeming2002; Brulez et al. Reference Brulez, Pike, Reynolds, Deeming and Reynolds2015; Reed and Clark Reference Reed and Clark2011). Whilst there has been an historical under-appreciation of avian sensory abilities (e.g. as noted by Birkhead Reference Birkhead2012; Brulez et al. Reference Brulez, Pike, Reynolds, Deeming and Reynolds2015; Caro et al. Reference Caro, Balthazart and Bonadonna2015), recent studies have demonstrated a range of sensory cues available to birds (e.g. Balthazart and Taziaux Reference Balthazart and Taziaux2009; Campagna et al. Reference Campagna, Mardon, Celerier and Bonadonna2012; Caro et al. Reference Caro, Balthazart and Bonadonna2015; Grieves et al. Reference Grieves, Gilles, Cuthill, Székely, Macdougall-Shackleton and Caspers2022; Mariette Reference Mariette2024; Ziolkowski et al. Reference Ziolkowski, Gracheva and Bagriantsev2022) that might be used by parents to dynamically assess current clutch/egg viability (see Table 1 for a summary). However, since it is known that some birds continue to incubate non-viable eggs beyond normal incubation periods (e.g. Briskie and Sealy Reference Briskie and Sealy1988), it is likely that the ability to use sensory cues in this way, and/or the propensity to abandon eggs, varies between species.
Summary of sensory cues available to parental bird(s) that may be used to examine egg viability status during incubation*
* Whilst here we mainly present summary examples specifically related to embryo–parent and embryo–embryo interactions during incubation, we also present some examples outside the context of incubation that further demonstrate birds’ sensory capabilities in non-incubation contexts, but in turn could also be used in an incubation context (although not directly proven).
Here, we first highlight current issues and inconsistencies with monitoring egg abandonment in avian research projects and monitoring schemes more broadly, via a systematic review of the literature. We then demonstrate how these issues can impact our understanding of the drivers of hatching failure in a conservation context, using data from on-going nest monitoring schemes for the UK and Ireland Eurasian Curlew Numenius arquata populations. Finally, we offer recommendations for standardising the long-term monitoring of avian egg abandonment and hatching failure to produce more accurate and informative data, which will likely also apply to the conservation management of other oviparous taxa with parental care (e.g. Pythonidae).
Issues and inconsistencies in the collection of egg abandonment data
To gain an overview of how egg abandonment is defined and measured across published bird studies, we conducted a systematic review by searching the database Web of Science on 24 October 2024 for the terms “egg”, “abandon*”, and “bird”. We excluded the terms “parasite” and “parasitism” from our search, because while egg abandonment is perhaps most intensively studied in species that experience brood parasitism, the focus of our study is abandonment by parents of their own eggs, which specifically requires better understanding from a conservation management perspective. After restricting the search to primary literature only (i.e. excluding review papers) and removing any duplicates, this returned 141 articles. Of these, 30 were removed because they were not published in English and translation was not possible, and a further 36 were excluded because it was evident from screening the abstract or main text that they were not relevant (e.g. they did not directly study or collect data on abandoned eggs/clutches or focused on brood parasitism despite our exclusion of related terms in the initial search). We read the remaining 75 papers in detail to identify: (1) what definition (if any) of egg abandonment was used; (2) what causes (if any) of egg abandonment were cited; (3) what rates of egg abandonment were reported. Full details of all retained studies are provided in Supplementary material data 1.
Of the 75 relevant studies, 43 (57%) did not provide a definition of egg abandonment for their study or describe how it was determined. The remaining 32 papers provided definitions of varying clarity, with 25 (33%) providing a definition that we considered to be sufficiently specific to enable reproducibility (e.g. the presence of intact but cold eggs in nest), but only 13 of these (17% of all identified studies) provided a conclusive timeframe for abandonment as part of their definition, e.g. “if we did not observe the breeding pair near the nest (or signs of activity) during two consecutive visits (3–4 days)”.
Most (64/75; 85%) of the studies identified provided some indication of the definite or presumed cause(s) of egg abandonment. These fell into 10 broad categories (see Table S1 in Supplementary material 1). Several studies posited multiple different causes of abandonment. Confidence in these cited causes ranged from relatively low (e.g. causes were assumed based on anecdotal evidence or findings of other studies) to high (causes were tested directly by the study). Overall, the main drivers of egg abandonment appeared to be human disturbance (31% of studies; predominantly due to farming practices and research/conservation interventions), climatic factors (19% of studies), disturbance by predators (17% of studies), and food availability (17% of studies). Notably, four studies confidently reported egg failure prior to abandonment, i.e. lack of embryo development was likely to be the reason the parents left the eggs.
Potential impacts of inaccurate abandonment records on breeding monitoring
Of the 75 studies identified above, 43 provided data on the proportion of clutches that were abandoned during the study period. However, one paper reported data for nine different species/populations, making 51 records in total. The mean rate of egg abandonment across all records was 20%, but rates varied from 1% to 100%, with a median of 12%. This represents a significant loss of reproductive potential, particularly given the increasing evidence that most wild bird eggs are fertilised and therefore contain individuals that died early (Evans and Postma Reference Evans and Postma2025). If egg abandonment occurred after the embryo(s) died – which it did in several studies – these losses will also have a significant impact on our understanding of hatching failure rates and drivers.
To assess the potential impact of incorrect abandonment records on quantitative monitoring of reproductive outcomes, we analysed egg outcomes recorded by various nest-monitoring teams working with wild populations of Eurasian Curlew (henceforth Curlew) across the UK and Ireland, using data collected for another study. The Curlew is a charismatic wader with around 19–27% of the global breeding population residing in the UK (Brown et al. Reference Brown, Wilson, Douglas, Thompson, Foster and McCulloch2015). Globally, the Curlew is currently classified as “Near Threatened” (BirdLife International 2017) and UK and Ireland breeding populations have seen significant declines ranging from 32% to 96% over the last several decades (Heywood et al. Reference Heywood, Massimino, Balmer, Kelly, Noble and Pearce-Higgins2023; O’Donoghue et al. Reference O’Donoghue2019). Consequently, Curlew are currently Red Listed and have been proposed as the most urgent bird conservation priority in the UK (Brown et al. Reference Brown, Wilson, Douglas, Thompson, Foster and McCulloch2015; Stanbury et al. Reference Stanbury, Eaton, Aebischer, Balmer, Brown and Douse2021). Low global Curlew productivity appears to be largely driven by nest predation (e.g. Grant et al. Reference Grant, Orsman, Easton, Lodge, Smith and Thompson1999; Roodbergen et al. Reference Roodbergen, Van Der Werf and Hötker2012; Valkama et al. Reference Valkama, Currie and Korpimäki1999; Zielonka et al. Reference Zielonka, Hawkes, Jones, Burnside and Dolman2019), and other factors such as agricultural land use, forestry, and climate warming (Franks et al. Reference Franks, Douglas, Gillings and Pearce-Higgins2017), but little is known about the incidence and non-predation-related drivers of hatching failure across UK and Ireland populations.
Combining data from 1,428 eggs, from 381 naturally incubated clutches laid in various locations around the UK and Ireland, across three breeding seasons (2022–2024), we calculated the percentage of all Curlew eggs recorded as abandoned to be 5.5% (although this was variable across years: 2.5% in 2022, 4.0% in 2023, and 8.0% in 2024; see Supplementary material 1 and Supplementary material data 2 for calculation methods and data). Hatching failure rate, excluding eggs classified as abandoned, predated, damaged, and “unknown” (i.e. when egg outcome was unclear), was calculated as 12.9% (11.6% in 2022, 13.9% in 2023, and 13.1% in 2024; see Supplementary material 1 and Supplementary material data 2 for calculation methods and data). However, if we were to assume that all abandoned eggs died before abandonment (i.e. abandonment was the consequence of embryo mortality, not the cause), recalculation of our hatching failure rates including abandoned eggs (i.e. the total number of “failed” plus “abandoned” eggs, divided by the total number of “hatched”, “failed”, and “abandoned” eggs) results in a 1.4-fold increase in our estimate of mean Curlew hatching failure rates to 18.6% (14.1% in 2022, 17.9% in 2023, and 21.7% in 2024).
While it is unlikely that all records of egg abandonment are incorrect, our experience of receiving egg samples and accompanying data has revealed that at least some degree of error is probable. For example, we frequently receive unhatched egg samples which are recorded as “abandoned”, but the recorded abandonment date is after the expected end of incubation (i.e. the parents left the clutch after it failed to hatch). Abandonment was therefore not the cause of failure, and these eggs should be included as sampling units in calculations of hatching failure rates. Confidence in egg abandonment dates is clearly dependent on nest monitoring effort, with higher frequencies of nest visits being more accurate. In many nest-monitoring programmes, failed eggs left in the nest after the end of incubation may be incorrectly recorded as “abandoned” due to a lack of information on lay date and/or expected hatch date.
By examining the embryo developmental stages of all intact unhatched Curlew eggs using previously developed egg dissection and fluorescent microscopy methods (Assersohn et al. Reference Assersohn, Marshall, Morland, Brekke and Hemmings2021; Birkhead et al. Reference Birkhead, Hall, Schut and Hemmings2008; Morland et al. Reference Morland, Patel, Santure, Brekke and Hemmings2024), we observed that while embryo development stages were fairly consistent within the majority of reported abandoned clutches, 5/25 (20%) clutches reported as abandoned had pronounced variation in embryonic development stages across eggs. Estimated embryonic developmental differences between the least and most developed embryos within each clutch ranged from at least 3 days to 11 days, which are likely conservative estimates of duration since timing is based on Hamburger and Hamilton’s (Reference Hamburger and Hamilton1951) embryo staging series for domestic Chicken Gallus gallus domesticus, which has a shorter mean incubation period than the Curlew (~21 days vs ~28 days, respectively (see Table S2 in Supplementary material 1, and also Supplementary material 1 and Supplementary material data 3 for further information on methods and data). Curlews have semi-synchronised hatching, with intervals ranging from several hours to up to two days (De Jong et al. Reference De Jong, Bocher, Brown, Franks, Gerritsen and Meyer2021), but the wider developmental stage variation exhibited in several reported abandoned clutches suggests at least some eggs experienced embryo mortality before egg abandonment occurred. Our evidence suggests that 20% of reportedly “abandoned” clutches contained eggs that experienced embryo mortality which was not caused by abandonment. If the parent(s) was able to detect embryo death/egg inviability (see Table 1 for potential mechanisms), it is possible that this subsequently stimulated egg abandonment. If it is possible to incorporate embryo staging into nest-monitoring programmes, variation in the developmental stages of dead embryos within a clutch may therefore provide useful insights into the likelihood of abandonment versus other factors as the cause of failure.
Conclusions and Recommendations
There are many well-established nest-monitoring programmes across the world, such as the British Trust for Ornithology (BTO) Nest Record Scheme, that collate records of breeding success across species. In general, these programmes record incidences of nest predation, damage, hatching failure, and abandonment, but what constitutes abandonment is not explicitly clear; for example, the revised 2003 BTO Nest Record Scheme Handbook (Crick et al. Reference Crick, Dudley, Glue, Beaven and Leech2003) provides no description of how to verify abandonment. We have shown, across a range of bird breeding studies, that there are significant inconsistencies in how egg abandonment is defined and monitored. Using empirical data, we show that eggs may be inaccurately recorded as abandoned due to ambiguous criteria and/or incomplete information on incubation dates/parental behaviour, skewing our understanding of how different factors contribute to overall rates of hatching failure.
In the ideal (but unlikely) scenario that time and resources are not limited, direct observations of parental behaviour, either via frequent nest visits or (even better) remote camera and/or nest temperature monitoring, provide the most irrefutable evidence of abandonment or “true” hatching failure. However, monitoring effort is usually limited, and uncertainty exists in most cases. As a result, research studies and conservation monitoring programmes differ in their approach to recording the causes of egg loss/hatching failure, potentially complicating our overall understanding of the drivers of failure across species.
Examining the developmental stage of dead embryos to assess whether they died before their predicted abandonment date is one potential way to overcome the problem of limited information on incubation timing/parental care. Methods for doing this are relatively straightforward; however, because embryo staging series have not been produced for most species, the approach relies primarily on comparison with the domestic Chicken (Hamburger and Hamilton Reference Hamburger and Hamilton1951) (precocial species) or Zebra Finch Taeniopygia guttata (Murray et al. Reference Murray, Varian-Ramos, Welch and Saha2013) (altricial species). While Zebra Finches probably provide a reliable comparison for many passerine species (Hemmings and Birkhead Reference Hemmings and Birkhead2016), caution should be taken when using this approach for precocial species, since the incubation periods and late-stage developmental trajectories of many precocial birds diverge somewhat from that of the Chicken (Cooney et al. Reference Cooney, Sheard, Clark, Healy, Liker and Street2020). This may lead to some degree of error in the estimation of embryo death timings based on developmental stages.
Finally, in situations where monitoring effort is limited and assessment of embryo developmental stages is not possible, we recommend that the resulting uncertainty in egg fates is clearly documented, and that studies report hatching failure rates both including and excluding presumed abandoned eggs, highlighting the range within which the “true” value must lie and enabling clearer assessment of the potential error in these estimations. As we found evidence that 20% of assumed “abandoned” Curlew clutches had experienced embryo mortality prior to abandonment, it is likely that the true hatching failure rate in this species will fall much closer to the traditional calculation of hatching failure (i.e. excluding abandoned eggs). However, where the true hatching failure rates lie on this spectrum is likely to vary between species, depending on factors such as propensity to abandon (e.g. see Hanssen et al. Reference Hanssen, Erikstad, Sandvik, Tveraa and Bustnes2023) and/or ability to use sensory cues (see Table 1) to examine current clutch viability throughout incubation. Future studies should therefore seek to obtain better understanding of how egg abandonment behaviour varies across species and the degree to which incubating birds can detect and respond to egg viability cues.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0959270925100221.
Acknowledgements
We greatly appreciate the Curlew Recovery Partnership, in particular Russell Wynn and Ryan Burrell, for facilitating communications to curlew nest monitoring fieldworkers across the UK and Ireland, and the following organisations and people for providing samples and data for analysis: Royal Society for the Protection of Birds (RSPB), Wildfowl and Wetlands Trust (WWT), Game and Wildlife Conservation Trust (GWCT), Banbury Ornithological Society, Nidderdale Moorland Group, Natural England, Upper Thames Wader Group, Grosvenor Family Office and Rural Estates, Shropshire Ornithological Society, Curlew Country, British Trust for Ornithology (BTO), Birdwatch Ireland, Pensthorpe Conservation Trust, Curlew Recovery South Lakes, National Parks and Wildlife Service Ireland, The Duke of Norfolk Estate, Curlew Recovery West Cumbria, Bannau Brycheiniog National Park, Nidderdale Area of Outstanding Natural Beauty (AONB) group, Hartwith Curlew Project, Denbighshire Government, Curlew Connections Wales Project, Dave Parish, Louise de Raad, Marlies Nicolai, Amy Burns, Luíse O’Donovan, Sarah West, Tracy Johnson, Anne Cotton, Rob Foster, Leo Smith, Hattie Jones, Stephen Dodd, Harry Ewing, Sam Franks, Elli Rivers, Kathryn Finney, Chrissie Kelley, Tony Sainsbury, Kane Brides, Eric Heath, Geoff Hilton, Scott Petrek, Barny Sykes, Susannah Bleakley, Tonia Armer, Heather Sykes, Joshua Burge, Paul Parmenter, Hubert Servignat, Philip Connolly, Tom Orde-Powlett, Charlie Mellor, William Costa, Max Wright, Rosa Lopez, Rebecca Lee, Lynda Donaldson, Chris Batey, Ken Moore, Nancy Reed, Ian Cole, Emily Hewitson, Lucy Foster, Sam McCready, Thijs Claes, Tanya Grigg, Alex Large, Dan Gordon-Lee, Paul Noyes, Sophie Common, Lucy Rivett, Elliott Simpson-Brown, Dan Gornall, Chris Heward, Pia Key, Mike Pollard, Jan Guilbride, David Hardwick, Thalia Sparke, George Purcell, Kelly Powell, Rachel Taylor, Verity Picken, Róisín Normanly, Andrew King, Gerry Hanson, Charlie Calvert, Will Price, Philip Miller, Samantha Kenyon, Katie Appleby, Seth Howell, Caleb Stradling and all others involved in on-going conservation efforts. NH was funded by a Royal Society Dorothy Hodgkin Fellowship DHF160200 and JET was supported by a Royal Society Enhanced Research Expenses Grant awarded as part of NH’s Fellowship. For the purpose of open access, the authors have applied a Creative Commons Attribution (CC BY) licence to any Author Accepted Manuscript version arising from this submission.
