Because of both the energy costs and the slowness of the reactions of the nitrogenase complex compared with those involving some form of combined nitrogen (oxidised or reduced), we argue that the evolution of nitrogen-fixing organisms required an environment which was very limited in combined nitrogen. This is thought to have occurred after phototrophy evolved, but before water was used as a hydrogen donor (and therefore oxygen was present in the atmosphere). After oxygenic photosynthesis evolved, the need for a high level of biological nitrogen-fixation remained, since abiotic inputs were insufficient to keep pace with the rapidly evolving biomass (flora and fauna). Symbiotic fixation probably first evolved in the form of casual associations between cyanobacteria and most other groups of plants. By inhabiting the sporophytic generation of evolving land plants (cycads in particular), protection against nitrogenase-inactivating oxygen and a more desiccating environment was achieved simultaneously.

We envisage nodulated plants arising by the transfer of nif genes into tumour-forming bacteria. In the case of legumes, these would be ancestors of extant agrobacteria, which gain entry into their hosts via wounds. Co-evolution of symbionts from nitrogen-fixing tumours has taken several routes, leading to extant nodules differing in mode of infection, structure and physiology. Evolution towards optimisation of oxygen usage is continuing.

Nitrogen-fixing symbiosis in animal systems is only advantageous in specialised ecological niches in which wood is the sole dietary intake. In the case of shipworms, the symbiosis has many of the advanced features associated with nitrogen fixing root nodules.

A wide variety of plant-microbe nitrogen-fixing symbioses which include cyanobacteria as the nitrogenfixing partner exist. While some information has been gathered on the biochemical changes in the cyanobacterium upon entering into symbiosis, very little is known about the accompanying changes at the genetic level. Much of our present knowledge of the organisation and control of expression of nitrogenfixation (nif) genes is derived from studies of the free-living diazotroph Klebsiella pneumoniae. This organism thus provides a model system and source of experimental material for the genetic analysis of symbiotic nitrogen fixation. We describe the use of cloned K. pneumoniae genes for nitrogen fixation and its regulation in the genetic analysis' of nitrogen fixation in cyanobacteria which can enter into symbiosis with plants. These studies reveal some dissimilarities in the organisation of nif genes and raise questions as to the genetic control of nitrogen fixation in symbiosis.

The part played by lichens in the nitrogen economy of natural ecosystems is reviewed. The ability of some lichens to fix nitrogen and the consequent likelihood of an enhanced contribution is discussed, leading to the appreciation that release of nitrogen compounds from their thalli is a frequent characteristic of their growth and development, and of some nutritional benefit to associated plants.

Data on the contribution by the normal processes of growth, death and decay, to the combined nitrogen supply of those specialised habitats where lichens predominate or are frequent, are given. Such habitats are arctic and subarctic forest and tundra, and the Douglas fir forests of northwest U.S.A. The methods used for the estimation of the contribution of the lichens are outlined and details are given of recent and current studies of nitrogen movements by direct leaching during normal growth of nitrogen-fixing terricolous lichens in Finland and southwest Scotland. The latter study uses a laboratory apparatus to simulate field conditions in order to permit the use of atmospheres enriched with the heavy isotope of nitrogen, 15N.

Alnus rubra growing in the field in Scotland is nodulated effectively by local strains of Frankia. Strains which have been isolated from Alnus rubra at Lennox Forest show different growth medium requirements and colony morphology compared with isolates from Alnus glutinosa and several different strains have been isolated from the same locality. Preliminary evidence suggests that some spore negative, northwest American Frankia strains may be more effective than local spore positive crushed nodule inoculum for nitrogen-fixation in Alnus rubra.

On the moderately fertile clay soil of Lennox Forest, a mixed plantation of alternate Picea silchensis and Alnus rubra showed no improvement in growth of spruce compared with pure spruce plots. However, the presence of alder increased upper soil nitrogen status by 585 kg ha ' which approximates the standard 150 kg N ha 'of fertiliser nitrogen applied in practice to nitrogen deficient stands at about 5-year intervals. Although the alder had penetrated the subsoil, there was no apparent effect on spruce rooting depth in the mixed stands in this high clay soil. In addition to nitrogen content, the total phosphorus of the upper soil horizons was improved in the mixed plots by an estimated 3–6 kg ha -1 y 1 and it is suggested that this phosphorus may be brought from the subsoil by the deeper rooting alder and deposited on the surface in its litter.

Although improved growth of spruce in mixture with Alnus rubra is only likely where pure spruce stands are stressed for nutrients, the potential benefits of symbiotic nitrogen-fixation in silvicultural practice make it desirable to investigate other species and provenances of Alnus suited to British conditions, to achieve maximum symbiotic fixation of nitrogen by selection and inoculation with superior strains of Frankia and to include such plants in trials of mixtures on sites where nitrogen-availability may be critical.

Interactions amongst plants and different endophytes are prevalent in soils deficient in both nitrogen and phosphorus. Several systems are now recognised, combining infections by both fungi and prokaryotes. Symbiotic associations are ancient and reflect the requirements for the maximum uptake of nitrogen and phosphorus in plant nutrition.

Sequential observations were made of the effects on root development of inoculating Betula pendula seedlings during propagation with sheathing mycorrhizal fungi, Amanita muscaria, Hebeloma sacchariolens or Paxillus involutus. After propagation, the seedlings were grown for two seasons in four different soils (two mineral soils and two peats) which were not sterilised.

While all inoculated seedlings had similar numbers of mycorrhizas at planting, numbers subsequently increased most rapidly with P. involutus and least rapidly with A. muscaria. From the end of the first season onwards, the uninoculated controls had as many mycorrhizas as seedlings inoculated with either A. muscaria or H. sacchariolens. These numbers were, however, usually exceeded on plants inoculated with P. involutus.

The mycorrhizas developing on the controls, and on seedlings inoculated with A. muscaria, were attributable after planting into the unsterile soils, to naturally occurring soil-borne inocula of species of Cenococcum, Hebeloma, Inocybe, Laccaria, Paxillus and Thelephora etc. Those developing on seedlings inoculated with P. involutus were all attributable to this inoculant fungus in the first year but, in the second year, the proportions decreased to 19—14%. With H. sacchariolens, at least 90% of the mycorrhizas in the two mineral soils and one of the peats (a sedge peat) were attributable in years 1 and 2 to the inoculant fungus. In the other peat (Sphagnum peat), the proportion of mycorrhizas attributable to H. sacchariolens decreased to 25% by the end of the first year and totally disappeared in the second.

The continued ability of H. sacchariolensto form mycorrhizas in a range of unsterile soils, to the virtual exclusion of mycorrhizas attributable to other fungi, is thought to be characteristic of fungi that occur early in mycorrhizal successions. In contrast,A. muscaria(a late-stage fungus), was unable to form mycorrhizas after inoculated seedlings were transferred from controlled axenic conditions to unsterilised field soils.

Two arbiters of root growth were measured—numbers of roots and their dry weights. Whereas inoculations with H. sacchariolens significantly decreased numbers in the latter part of the first season, they either had no effects on dry weights as in the mineral soils or significantly increased them as in the peats. In the second season, there was further evidence, sometimes contradictory, showing that inoculations can appreciably alter the structure of developing root systems.

Some aspects of the role of the ericoid mycorrhizal symbiosis in the ecology and physiology of ericaceous plants are described. Mycorrhizal infection leads to enhancement of plant nitrogen content and an experimental analysis of the basis of this effect is reported. In addition to improving the efficiency of ammonium absorption at low concentrations, the mycorrhizal endophyte utilises amino acids, peptides and proteins as nitrogen substrates for growth. These are the predominant nitrogen sources in organic heathland soil. It is suggested that the success of ericaceous plants in such soils may arise through the capacity of the mycorrhizal fungus to provide its host with access to this nutrient resource. A model is described in which absorption of ammonium and amino nitrogen leads to soil acidification, increased acid protease activity and improved vigour of the ericaceous plants.