Worldwide, there are many species of small birds that live in trees and mainly eat invertebrates. However, finding out exactly what invertebrates they are eating has always been tricky and is not well known, even for very common and well-studied species like the Blue Tit, which cheerfully hops around the gardens of millions of people throughout Europe. The reason for this is that the birds themselves are small, their prey is even smaller and difficult to identify in the field, and the predator-prey interaction is often happening very fast and high in the canopy, making observation almost impossible.
So, how has identification of prey been achieved in the past? For static and accessible chicks in nestboxes this has proven easier, with the use of neck collars and cameras in the nest giving good coverage. For adults, the only way has been euthanising the birds and dissecting the gut and gizzard contents, and this is not ideal – it is very destructive, excludes soft bodied dietary items from being observed and only allows identification to a low resolution. In this study, we show how large-scale faecal metabarcoding (genetically identifying prey contained within faecal samples) can overcome these challenges, identifying prey to excellent resolution across time and space in a completely non-invasive and non-destructive way. In this post, I am going to discuss the biological and natural history advances that this study has provided.
Firstly, it is interesting to note how diverse the diet of Scottish blue tits is in early spring. We identified over 400 different invertebrate prey items from nearly 800 samples at a time of year when invertebrate resources are low because leaves are not yet out. The fact that most of these were rare, with almost half only recorded from a single sample each, suggests that blue tits have a wide dietary range and are opportunistically taking advantage of many food resources as and when they find them rather than solely focussing on particular items. This is supported by there being an average of five, and up to twenty, different prey items per sample. Some prey items were favoured, with 15 species present in over fifty samples each, the most common of which (the micromoth Argyresthia goedartella) present in over a third of samples. Most of these common species, including the caterpillar of this species, were associated with resources available in early spring, such as birch catkins, but I will return to this later. Moths in general were both the commonest and most species rich prey items, with at least one of over 100 identified species present in three quarters of samples. Other common prey orders included aphids, flies, beetles, gall wasps and spiders.
Collecting faecal samples throughout early spring from a transect of 40 (now increased to 44) field sites (which cover 2° of latitude, 400m in elevation and many different woodland types) allowed us to assess how blue tit diet changed along environmental gradients and as spring progressed. Dietary variation could either be due to local resource availability or preferences, and seasonal variation in diet has implications for time-sensitive life events (such as breeding or moving) while spatial variation could affect population density or productivity. We show that dietary richness increases as spring advances but does not change significantly with geography or habitat, whereas dietary composition exhibits significant turnover along temporal and spatial gradients. Such insights into how a generalist insectivores diet varies over space and time are very rare indeed. The high resolution of the prey data also allowed us to show that moths increase in prevalence in the diet with increased latitude and elevation and aphids show a steep increase in prevalence in the diet as spring progresses, presumably as they emerge to feed on growing buds and fresh leaves.
One particularly intriguing dietary occurrence is winter moth, a species whose caterpillars are known to be a crucially important food source for feeding nestlings later in the spring, but whose presence in the diet this early in spring was unexpected. As their occurrence in the diet increased throughout the sampling period, from a 2% chance thirty days before egg laying to a 17% chance at the point of egg laying, we assume that the life stage being preyed upon is early instar caterpillars. We also forward the possibility that feeding on these caterpillars provides a direct cue to the birds to start laying eggs, as how the birds time their breeding to coincide with the late spring winter moth caterpillar peak is an area of intense research and speculation. Although we cannot provide any direct evidence of this, it would provide a very direct and probably very reliable cue and offers an interesting new avenue for further work.
As mentioned previously, much of what we knew about blue tit diet was based upon microscope identification of physical prey remains encountered within the guts of euthanised birds, with the most comprehensive studies being carried out in southern English oak woods in the middle of the last century. One of my personal highlights was the ability to compare our genetic prey dataset from Scotland today with those physical datasets from England 60 years ago. Rather incredibly, some of the common taxa identified were identical, including the springtail Entomobrya nivalis, gall wasps of the genus Andricus, and several species of fly larvae feeding on emerging tree buds. Not only does this validate our method and emphasise the widespread nature of these dietary items, it is also testament to the immense skill and patience of those with microscopes and tiny body parts to identify!
However, many of the other species that we identified were previously unknown dietary items. This is due in part to the extra identification resolution that we were able to utilise with these new techniques allied with our larger sample size, but also in part due to the differing habitats studied. Half of the ten commonest taxa we identified were exclusively or primarily associated with birch trees, which is by far the commonest tree across our study region but is far less common further south in England. This could highlight an adaptability of insectivorous birds to flexibly adjust diet according to local food sources and raises the possibility that one reason we didn’t find a significant effect of habitat on diet is due to the widespread nature of birch across our sites. It will therefore prove interesting to contrast diet with regions with differing locally abundant trees if this method gains more widespread use.
More specifically, many of the commonest dietary items found in this study, particularly the commonest moth caterpillars (Figure below), are associated with birch (and alder) catkins. This explains previous observations of tits foraging on catkins at this time of year, but rather than eating pollen or plant material as was hypothesised, it appears that they are opening the catkins to prey on the caterpillars living and feeding inside them. Equally, our dietary results are consistent with observations of tits feeding around birch and sycamore buds on freshly emerging aphids, with three out of four of the commonest aphid species associated with these. The ability of the results from this method to tie in behavioural observations, physical studies, and reflect the local habitat creates an accurate and believable picture of the diet of blue tits across Scotland at this time of year and validates metabarcoding as a technique to better understand the diets of insectivorous birds.
To find out more, read the full paper here: https://doi.org/10.1111/mec.15394
Dr Jack Shutt,
Postdoctoral Research Associate, Manchester Metropolitan University.