Open Access

David Brooks, John Bater, H Jones, P. A. Shah

• 1.1 Introduction 1.1.1 Aim of project .To compare the abundance of invertebrate and weed seed food resources available to birds on orzanic and conventional farmland. The objective of the studv was to assess accurately the likely benefit of these farming systems to birds feeding on farrn'land, by sampling invertebrates and weed seeds. 1.1.2 Factors implicit to achieving the project aims Variation between farms within one system could influence' invertebrate or weed seed abundance and bias results, To minimise such effects and provide results 'representative of the farming systems as a whole, sampling was based on an extensive approach; farms were sampled in groups. The inference that either of the farming systems is beneficial to feeding birds is dependent on: (i) prior knowledge of the relevance of a particular invertebrate or seed as a food-source: evidence (ii) that this foodsource is present in sufficient abundance and (iii) that the food-source is readily accessible, The methodology described below was refined to address these criteria. 1.1.3 Methodology Sampling concentrated on cereal crops, with an additional comparison of organic grass ley fields at a limited number of sites. Sampling initially consisted of sucking invertebrates from the crop using a vacuum insect net and extraction from soil cores. To aid the interpretation of results from this sampling, it was decided that more information was required on the diet of birds. This was achieved by analysing faecal sacs for undigested fragments of invertebrates that therefore represented a dietary component. Skylark chick faecal sacs were chosen for analysis as this was the key species for the intensive ornithological studies and samples could be taken during routine fieldwork. As a result of this study, the main invertebrate sampling technique was changed to pitfall traps, since this was a superior method for assessing those invertebrates found to be important food-sources. It was also anticipated that pitfall trapping would provide more accurate estimates of invertebrate availability, with greater numbers per sample, than the previous two techniques. Studies of weed seed food resources consisted of field surveys using a quadrat to assess the presence and abundance of species, and the use of a small hand held suction machine to suck seeds from post-harvest stubble. The interpretation of the results emphasised the aspects of the ecology of species known to be food-sources that might influence their availability to birds. 1.2 Soil core and vacuum samples 1.2.1 Significantly more dipteran immature stages and Coleoptera were found in soil cores on organic grass ley fields and significantly more earthworms on organic cereal fields than conventional cereal fields. Earthworms and dipteran larvae such as tipulids are known to be important food-sources for birds that specialize in soil invertebrates. 1.2.2 Total numbers of invertebrates trapped by both methods did not differ significantly between the two farming systems. Significantly more invertebrates were trapped, however, by both methods on organic grass ley fields than either conventional or organic cereals. 1.2.3 Significantly more Staphylinidae tCol.), especially the species Tachyporus hypnorum, were found on conventional fields, The relevance of this species as a food-source is, however, doubtful. 1.2.4 The weevil Sitona lineatus and the carabid Demetrias atricapillus were found in significantly greater numbers on organic fields. The former may constitute a food-source for skylarks, which have been shown to feed on this insect under laboratory conditions. 1.3 Faecal sac analysis 1.3.1 Carabid beetles were an important component of skylark chick diet, forming 47 % of identifications. In some cases it was possible to identify the species present. 1.3.2 Coleoptera, other than carabids, from the families Elateridae, Curculionidae, Chrysomelidae and Staphylinidae were identified as food-sources. 1.3.3 Spiders and tipul ids were also important components. 1.3.4 Reservations are expressed that the technique may under-represent soft-bodied invertebrates, which are susceptible to complete digestion by skylark chicks. 1.4 Pitfall trap samples 1.4.1 Twelve key species of carabid beetles were analysed. of which five cornmon species were trapped in significantly greater numbers on organic farms. These were Pterostichus melanarius (the dominant species captured), Pterostichus madidus, Harpalus affinis, Harpalus rufipes and Nebria brevi collis. The other species analysed showed no significant variance between farming systems. 1.4.2 Williams' Index of Diversity was significantly greater for conventional fields, although caution is expressed over the reliability of this result due to the small sample size of species. 1.5 Botanical studies 1.5.1 The abundance of weed plants in quadrats was significantly greater on organic fields. 1.5.2 The abundance of weed seeds was not significantly different between farming systems. However, the relative proportions of monocotyledonous and dicotyledonous seeds in samples differed between farming systems. A larger proportion of the seeds from organic fields were dicotyledonous and from conventional fields were monocotyledonous. 1.5.3 Preliminary examination of the size of plants and the number of seeds produced suggested that those on organic fields may have been nitrogen deficient. 1.5.4 Weed species were significantly more diverse on organic fields. although diversity has less relevance to bird feeding than abundance. 1.6 Proposals for future work 1.6.1 Replication of the pitfall trapping exercise in subsequent years would substantiate the trends established from the data of one season. It would also be beneficial 10 extend the range (If habitats sampled to take into account set-aside and other crops besides cereals. 1.6.2 More comprehensive information. on the diet of farmland birds in general. could he achieved by analysing the faecal sacs of a wider variety of species. 1.6.3 Greater integration of field studies on birds with invertebrate sampling would enhance the effectiveness of the latter as an indicator of diet. Areas of farmland frequently selected as feeding sites by birds could be sampled intensively for invertebrates and compared to other areas, selected at random. This would provide useful information on the invertebrates likely to be important as food-sources and the habitats that favour them. 1.6.4 Extending the range of farms sampled would provide more accurate results. 1.6.5 More work is required to-investigate the possible link between nitrogen deficiency in plams and organic systems, and its implications for the provision of bird food-sources, particularly for weed abundance and phytophagous insects.