emergence times for some species may result in significant under-recording of species present. Another complication is that species may be present at a site, but not breeding. A site which is not itself particularly rich in breeding species may appear to be better than it is if it is adjacent to other good sites. In Table 3:3 I have taken some account of this problem, indicating presence of a species with a plus, bracketing those species which are unlikely breeders at the site concerned. (Not all observers have consistently indicated breeding-status when sending in records.) A final complication has to do with what counts as a 'site'. Some localities, such as worked-out mineral extraction sites have more-or-less clear-cut boundaries, but this is less true of. for example, rivers, ditch-systems in grazing-marsh, or larger complexes of breeding sites such as Epping Forest. Data about the Essex rivers has been tabulated separately (Table 3:2). but for still-water or complex sites I have identified sites according to convention or convenience, without regard to strict comparability. 1. Complex Sites (a) Epping Forest Epping Forest has been subject to more intensive and long-term study than any other locality in Essex (see Appendix B). In the last century the Epping Forest area was notable for the presence of a number of scarce acid-bog species, but Henry Doubleday (1871) was already bemoaning the loss of habitat in the Forest. The early Transactions of The Essex Field Club and their successor. The Essex Naturalist, continued this concern through the 1880s and 1890s. At a meeting on February 25th, 1882, for example, one Dr. Cooke voiced alarm at the threat to Forest pools posed by the Conservators' policy of deep drainage. The Conservators were encouraged to 'consider the scientific Cockney as well as the Easter Monday Cockney in the preservation of Epping Forest" (Transactions, Vol. III, p.xiv). Subsequently, a report on a Forest ramble on June 20th, 1896 concluded: 'It was a subject of frequent remark, and was exceedingly disappointing to the naturalist to witness the slow extinction of the hog flora of the Forest, caused by the straight ditches cut through the bog land' (Essex Naturalist, Vol. IX. 1895/6. p.208). Very little of this habitat now remains. However, in recent years valuable conservation work has been done in the Forest, and small remaining areas of acid heathland and bog. especially on Sunshine and Deershelter plains, now look suitable for recolonisation by some currently absent species, (see above, under Acidic bog and Heathland Pools') Epping Forest still has, however, a pre-eminent position in the county for its dragonflies. This is largely because of the presence within the Forest of a number of ponds of various sizes and situations, together with large tracts of uncultivated terrestrial habitat. The Forest ponds have diverse origins. Some, such as Connaught Water and Baldwins Hill Pond, were formed by damming streams, whilst others, such as the Lost Pond and Strawberry Hill Pond, were formed by sand-and-gravel extraction, and yet others, such as the Wake Valley Pond, had their origins in the construction of the 'new' road through the Forest. Fairmead Pond was probably created to provide water for cattle grazing on the plain, and the Goldings Hill Pond may have originally been created as a horse-pond. 34