The objective of the Review has been to select according to established criteria of nature conservation value a series of sites which gives acceptable representation of all the more important features within the range of variation in natural and semi-natural ecosystems in Britain. The process of working towards this goal involves three distinct stages, though the second two are so interwoven that they tend to be thought of as one. These stages are as follows:

On the basis of the scientific record obtained in (i), the quality of each site is compared with that of other sites of similar type, so that there is a continual process of selection to identify the ' best' site within a related group.

This involves decision on which, how many and what extent of important sites should constitute the national series, and is the most difficult part of the whole process to rationalise satisfactorily, since it is essentially subjective, even when based on a consensus view. These three phases will now be considered in detail.

For a review of this kind to be possible, there is a need for adequate knowledge of the range of natural and semi-natural ecosystems in Britain, at least qualitatively, and preferably quantitatively, in terms of habitat and associated organisms. The analytical description and hierarchical classification of the whole range of ecosystems represents a taxonomic framework which gives a basis for identifying and recording the variety of features at any site, and enables one to compare different sites, discern geographical trends and differences, and select sites according to properly defined criteria. Ideally, one would wish to consider in this way all the components of an ecosystem, including the inorganic factors, human influence, plants and animals. In practice, such are the imperfections of our knowledge and capacity for descriptive analysis that we have at present to limit our attention largely to the vegetational component of the ecosystem. This is, by general consent, the most useful parameter by which to characterise the ecosystem, for it is an integrated expression of a complex of interacting environmental influences, and at the same time is the major determinant of the animal component.

A hierarchical classification of vegetation which uses environmental and life-form criteria to separate the classes of higher rank is more useful for present purposes than one which relies on purely floristic (i.e. phytosociological) criteria. Accordingly, a primary separation has been made into the seven main ecosystem groups of Coastlands; Woodlands; Lowland Grasslands, Heaths and Scrub; Open Waters; Peatlands; Upland Grasslands and Heaths; and Artificial Ecosystems. These are regarded as formations (= major habitat grouping or ecosystem type).

These are intergrading or even overlapping classes, with arbitrary and artificial boundaries drawn between them, and two or more of them may be represented within the same site. Peatlands, for instance, are well represented within upland grassland and heath sites. Nevertheless, this has proved to be a convenient and practical subdivision into classes of first rank. Separation into classes of second and third rank has been according to life form and structure of the vegetation, physiography, or edaphic criteria, as seemed most convenient. The ideal would be to produce final classes of lowest rank determined essentially by floristic criteria (vegetation or community types, noda, associations), but only for upland grasslands and heaths are phyto-sociological data sufficient to allow this degree of definition. The whole task has, in fact, been greatly handicapped by the lack of a standard countrywide description and classification of British vegetation types. Dominant, constant and characteristic plant species have been used to define the upland vegetation types. Even at this finer level of analysis, there are few if any real discontinuities in vegetational variation, and the limits between related vegetation types are mostly arbitrarily drawn. For the formations other than the upland grasslands and heaths, the description of vegetation has been more flexible, according to present availability of knowledge.

The units of vegetational classification of any rank may be used as the basis for ecosystem recording on any site, but the lower the rank of unit employed, the more detailed and accurate does the record become. In this vegetational taxonomy, the units of 'nodum' (community type) rank, even if only provisional, may be regarded as equivalent to species in classical taxonomy, i.e. they represent a concept of parallel utility. The diagnosis of vegetation types thus gives a means of identifying plant communities encountered subsequently in the field, and of recording these simply, thereby providing a rapid but detailed description of the range of variation. It is, however, essential to characterise ecosystems in terms of as many features as possible, and supplementary information has been added, as available. Inventories of plant species have been compiled for as many sites as possible; vascular plants and bryophytes have been recorded most fully, lichens less so, whilst fungi and algae have received uneven and inadequate study, and have often been omitted. Species which are rare, local or decreasing are indicated in the lists, as are members of important phyto- geographical elements.

The vegetation of open waters poses considerable problems of description, especially in terms of plant communities. Not only are there practical difficulties in examining and recording submerged vegetation, but the communities of aquatic plants are often open, heterogeneous and vertically stratified. Moreover, the algae, especially planktonic forms, are such an important component of the aquatic ecosystem that they must be considered as well, and the composition of the invertebrate fauna has also been used as an important criterion in the description and assessment of open water sites.

Apart from the open waters, it has not usually been possible to record animal communities as a parallel to vegetation types, and zoological records have consisted mainly of species inventories, with some indications of population size for certain vertebrates. Birds, mammals, reptiles and amphibia and freshwater fish have been fairly well recorded for most of the important sites. Recording of terrestrial and some aquatic invertebrates is still in a highly fragmentary and unsatisfactory state. Macrolepidoptera and Odonata are the best known taxonomic groups, but even here records are inadequate for many sites. Species lists for many groups of freshwater invertebrates, together with a measure of abundance, are available for most of the key open water sites. Specialist surveys of other invertebrate groups such as Curculionoidea (Coleoptera) and Arachnida, have given detailed knowledge for a limited number of sites, but patchy information of this kind is of mixed value, for it may lead to the overlooking or under-rating of unsurveyed but important sites. As with plants, efforts have been made to indicate occurrences of animal species which are rare, local, declining or of geographical importance. Knowledge of habitat and vegetation often makes it possible to predict the presence of characteristic animals with fair accuracy, and the various sections on fauna in vol. i follow an ecologically based treatment which indicates these relationships.

Climatic data of particular ecological relevance have been noted, especially rainfall, temperature and snow cover. The main geological and geomorphological features of sites have been recorded, with emphasis on lithology and especially the available lime content of the rock. Data on soil types and their chemical characteristics are available for some sites, but pedological information is rather uneven. Analyses of dissolved and particulate minerals are available for all open water sites. Biotic influences, especially those involving man, have been assessed and described whenever possible. Not only are past and present land use important because of their effects on the stability and present composition of the ecosystem, but some forms of land-use are also of intrinsic interest (e.g. coppice or water meadows). Even completely artificial sites with strong biological interest have been included in the survey. Detailed knowledge of the history of a site enhances its value, especially for teaching and research, so that relevant historical records have been included with site data whenever possible. The data for each defined area were entered on site record cards or check sheets and species record cards, boundaries selected and marked on a map, and an integrated statement prepared summarising the important environmental and biological features of the site.

The term ' site' has been used, for convenience, to refer to any surveyed area of land, large or small, for which prescribed boundaries can be drawn on a map to define a single geographical unit. There is a certain implication that these boundaries are clearly recognisable and self-defined on the ground, but this is often far from the case, especially in large continuous tracts of upland, where the limits to a site usually have to be arbitrarily defined.