Ensuring adequate isolation from pollen exchange among accessions is the single most important step in maintaining genetic integrity and reducing risk during regeneration.
Under natural conditions gene flow by pollen can be almost completely counterbalanced by selection against migrant alleles; thus, differentiation may occur in the presence of up to 50%
pollen flow, but the selection pressures have to be of a high order – up to 70% to keep the population differentiated in the presence of a pollen flow of 50%~60% [58]. The relationship between the opposing forces of migration and selection are complex, depending on the dominance relationships of individual loci and the inheritance of the characters involved
[59,60,61]
. In the absence of selection against migrant alleles, such as would be obtained in accession multiplication, even moderate rates of pollen contamination could cause significant alteration [58,62].
There are several ways to achieve adequate isolation, but these depend on the method of pollination (insect or wind) and the reproductive biology of the crop species. In general, isolation is obtained by: 1) spatial or temporal separation; 2) natural or artificial barriers; and 3) hand crossing accompanied by bagging of the inflorescences.
Spatial isolation studies for a number of wind and insect-pollinated species [63,64,65, 65a]
all shown significant (>5%) pollen dispersal by both wind and insects can occur over large distances, although this can vary by season, location and species, direction of wind and insect species and their preferences [39]. For example, most wind-dispersed pollen travels much further than animal-dispersed pollen [39]. For seed multiplication of allogamous cultivars, isolation distances in excess of 200m may be a statutory requirement [66,67]. When dealing with a large number of accessions requiring regeneration, isolation by distance is usually not practical. The outstanding feature of both wind and animal pollen travel is its leptokurtic distribution [68,69]. This distribution is influenced by the foraging strategy of the insect pollinators as related to within-plant and between-plant travel [70]. Few studies have investigated gene flow, as opposed to pollen flow, in wind-pollinated plants. In rye grass, Lolium perenne over 99% of the genes traveled less than 20m [68] and in beet, Beta vulgaris, all genes traveled less than 10m [71]. These values are of an order of magnitude less than the pollen travel observed in these species, and contrast sharply with insect pollinated plants in which gene travel usually exceeds pollen or pollinator travel [39].
Temporal isolation is sometimes practiced on perennial forage grasses [72]. Through cutting, the plots of different accessions are maintained at different stages of development, hence assuring that where crossing is minimal only distant accessions are at the same phenological stage, such as flowering or heading.
Pollen flow can be further interrupted and reduced by intervening barriers of other species, and this has led to the practice of growing isolation plots in standing crops of tall growing species, ie. sunflower, hemp. Several genebanks have exploited this method by establishing seed islands dispersed in crops of winter rye or winter wheat [73] or Brassica spp.
[74]. Isolation is achieved by 50m separation between the islands and several seed islands may be contained in one field. Each island contains populations of different wind and insect pollinated species, the latter relying on natural pollinators. The assumption is that the pollinating insects remain within a plot, and before approaching another plot they will be attracted by the flowers representing the island plants. Although this method does not ensure the avoidance of any gene flow between accessions, it is relatively simple and cheap and
thus a large number of accessions can be regenerated.
Other methods commonly use artificial (linen, paper bags and other materials) isolations for the period of flowering for wind pollinated species for controlling interaccession pollination [75,76,43]. The bags cover prior to anthesis a single head, 3~4 heads, in a method referred to as cluster bagging, or a whole plot with the intent to encourage cross-pollination and maintain genetic integrity among accessions [77,78]. Supplemental hand-pollination is sometimes required.
The use of screened cages with specially designed hives for honey, leafcutter or bumble bees or flies (hover, blow) for several insect-pollinated crop species is now common
[79,43,80,81,82,83]
. To save on resources, one accession each of different crop species may be planted under one net or nylon cage. Insect pollinators are released inside the cage at flowering time to enhance pollination within each accession and thus increase yield.
However, problems may occur when more than one accession is grown under a cage, and in come cases supplementary hand-pollination may be necessary, eg. wild species of tomato and eggplant. Also insect pollinators may show preference for one species. The USDA tested the ability of cages to restrict gene (pollen) flow [84]; compared seed produced by various races of honey bees [85]; and documented improvements in regeneration quality and quantity - for sunflower and carrot. Purpose-built pollen proof glass or polyethylene houses are effective, but expensive to operate and maintain [88]. An example can be found at the Institute of Grassland and Environmental Research, Wales which uses a rigorous regeneration scheme for all accessions (forage grasses and clovers) using elaborate crossing houses [89].
The use of isolation cages poses several problems. Aside from being expensive and laborious to construct, the shading that results affects plant growth and may distort characterization data, and pollinator foraging behaviour [90.43]. The environment inside a cage may be favourable for pest insects (aphids, mites); severe infestations may be frequent and often difficult to control even with repeated use of pesticides. Their use may result in negative pesticide-pollinator interactions and the need to reintroduce adequate numbers of pollinators. On the positive side, cages may result in the exclusion of pest, such as protecting Cucumis from beetle-transmitted bacterial wilt [91] or other insects that damage plants and/or fruits, which cause reduced seed yield as well as reduced seed vigour.
The ultimate effective isolation is controlled pollination by hand. During the early stages of seed regeneration hand-pollination can produce sufficient seeds, but with
expanded programs for breeding, testing and the handling of large collections with numerous accessions of each species and the need for effective population size to maintain integrity this becomes too laborious and time consuming. Furthermore, natural pollinators are more efficient and effective in producing quality germplasm than by hand.