Search

This research assesses the importance of using ecologically-similar rather than geographically-local seed in grassland restoration projects, with particular reference to herbivorous invertebrates, including pollinators. Seed from Lotus corniculatus L. (bird’s-foot trefoil) populations at nine sites across south-west England were collected to represent ecotypes potentially adapted to a range of soil types (calcareous loam, neutral loam and calcareous sand (referred to as ‘sand’)) and management regimes (grazed, cut with aftermath grazing (referred to as ‘cut’) and unmanaged). From October 2011 the ecotypes were planted within three different treatment soils (calcareous loam, neutral loam and calcareous sand (“sand”), and two management treatments (grazed [simulated] or unmanaged [neither cut or grazed]). Differences in plant morphology and phenology under these treatments were recorded at four-weekly intervals and immediately prior to harvest on 16th July 2012. Fresh and dry biomass were recorded and leaf-nitrogen and leaf-hydrogen cyanide (HCN) levels determined. Treatments were maintained and plants grown for a further 15 months [after harvest]. Bee preference for ecotypes grown under treatment combinations was also recorded during peak flowering periods of 2012 and 2013. Data were tested using Kruskall-wallis and ANOVA. A Generalized Linear Mixed Model (GLMM) was built to test all ecotype and treatment differences including interactions. A separate Non-linear Mixed Effects model (NLME) was created to investigate spatial autocorrelation between ecotype sites. The standard chosen in the models was Cockey Down (a calcareous loam, grazed ecotype) grown in matching treatments. The phenotypic traits retained were most pronounced in ecotypes from home-sites of sand soil type and cut management, which were considered to be the more stressed environments of the study, requiring rapid adaptation. By harvest, ecotypes from sand home-sites produced significantly greater number of stems per plant, greater leaflet number per main stem and lower HCN. Model results for sand ecotypes additionally identified delayed seed pod formation, increased hirsuteness and higher leaf-HCN compared to the standard. Significant differences found within ecotypes from cut home-sites included fewer stems per plant, fewer leaflets per main stem, more seeds per pod, greater leaf-HCN and shorter time to first flower. The model also found this ecotype to be less hirsute with fewer seed pods (in unmanaged treatment) than the standard. Ecotypic traits shown in plants from the less stressed home-sites included calcareous loam ecotypes having two clear flowering peaks in both years and highest leaf-HCN, and unmanaged ecotypes having lower leaf-HCN. Three significant interactions indicated additive character factor effects: neutral loam ecotypes grown in neutral loam treatment soil had earlier pod formation than the standard; and, sand ecotypes grown in sand treatment soil and unmanaged ecotypes receiving unmanaged treatment had greater flower number (over both years) than all other ecotypes, treatments and combinations. Results from the bee ecotype preference study showed no preference for ecotypes geographically close to the test foraging area. However significant differences were shown by bees in terms of ecotype preference, with avoidance of plants containing highest leaf-HCN. Plants grown in calcareous soil treatment were preferred which suggests nectar of plants are of most value to bees when grown in optimum [for L. corniculatus growth] soil. Ecotypic differences in herbivory defence [leaf-HCN and hirsuteness induced by home-site soil and management], would be of importance to receptor site invertebrate herbivores/pollinators. Pollinators could also find difficulties with the ecotypic differences in flowering asynchronicity. Both home-site soil type and management could also influence the viability of the plant population through reducing fecundity. Delayed seed pod formation in sand ecotypes (compared to the standard) indicates an adaptation to summer temperatures or low-nutrient availability. Calcareous loam ecotypes lack such adaptation and if introduced to a sand sites and could fail due to poorly timed germination, deep seed burial [from shifting sand] or poorly allocated energy. Cut ecotypes also produced significantly fewer seed pods (than the standard) with significantly greater seed numbers per pod suggesting an energy allocation adaptation to produce fewer, larger pods before defoliation rather than continuous pod formation throughout the season, a potentially critical adaptation for seed return. Findings from this study are of national relevance, and Natural England should adopt new recommendations on seed provenance in agri-environment schemes. Instead of recommending strictly geographically local seed, the management regime (particularly details of intensity and timing of management operations) should ideally be similar between the donor and receptor sites. Soil types, especially pH and clay/organic matter content should also be matched as far as possible as these were the greatest limiting factors within this study. Suitable donor sites may be local sites of similar habitat. If no such sites are available then recommendations from this study should be followed in seeking suitable sowing material. If seed suppliers are used, then they should provide greater detail on donor site conditions to aid land managers.

This thesis investigates a sequence of 17 cycles of coarse-grained syn-rift delta progradation in the Loreto basin, Baja California Sur. The Loreto basin is a Pliocene half-graben bounded by the 30 km long easterly dipping extensional Loreto fault, with 1.5 km throw. The Loreto basin uniquely provides a temporally constrained setting that couples preserved alluvial architecture with marine depositional geometries that enables these controls to be distinguished. The depositional architecture of each cycle is reconstructed by correlating 61 close-spaced (20 m to 320 m), detailed (1:5000) sections. Of the 17 progradational cycles, the lower 11 are particularly well exposed. Temporal constraints achieved through the dating of two sequence bounding tuff horizons, give an average duration for a single cycle of delta progradation to be 14±8.6 kyr., with most cycles being <10 kyr. This duration is too short to be explained by Pliocene eustatic variations (41 kyr). A consistency in measured palaeocurrents towards the NNE suggests that lobe switching was not a significant factor in controlling delta architecture. A facies framework is established that recognises both shoal-water and Gilbert-type delta deposition in the basin. The facies framework, coupled with reconstructed depositional architectures reveals that 7 of the 11 lower cycles, and at least one upper cycle (12) exhibit a palaeoseaward transition from shoal-water to Gilbert-type deltas. This nucleation of clinoforms with initially oblique, but increasingly sigmoidal geometries is characterised by a down-transport development of forests, which increase in height from 4 m to >26 m as they are traced basinward. Each cycle is capped by a shell bed that records drowning of the delta top. The geometry, and specifically the nucleation of clinoforms in single cycles of progradation requires a continued basinward increase in bathymetry. In one case this increase can be attributed to the topography of an underlying delta cycle. However for most of the cycles the increase in bathymetry can only be achieved if the rate of tectonically-controlled accommodation creation accelerates during their deposition.