THE NEED FOR CLEARLY DEFINED POPULATIONS

The experimentalist wishes to reduce the uncontrolled variation in an experiment so that it does not confound the effects of the experimental treatments. It is therefore crucial to correctly identify the species of weed present, and to ensure that the population of each species is clearly defined.

THE PROBLEM

Several species of weeds are difficult to distinguish at the vegetative stage from closely related species. This problem is often encountered with field populations of grasses, Matricaria, Veronica, Amaranthus etc. Even within a species, individual biotypes with significantly different tolerance of a herbicide may be widespread (Appendix 1 gives a few examples). Failure to define the weed population can have a crucial impact on the validity of experimental results as illustrated by the following examples;

Example 1). In a field trial which included chlortoluron and tralkoxydim to control Alopecurus myosuroides in cereals, one treatment killed 90% of the grasses in the plot. The facile conclusion was that the treatment gave 90% control of A. myosuroides. However, a closer examination of the grass weed population showed that it was composed mainly of A. myosuroides but with up to 20% of Lolium perenne and Poa annua. The true interpretation was that the tralkoxydim treatment gave 100% control of A.myosuroides but no control of the Poa. In the chlortoluron treatments the level of control achieved was strongly influenced by the tolerance level of the Alopecurus biotype(s) present and the interpretation would probably have been invalid if the presence of the usually more tolerant Lolium perenne had not been recognised.

Example 2). High throughput glasshouse screens tend to have very low numbers of plants per treatment. This makes it crucial to use a homogenous and clearly defined population of each species, since a small number of atypical plants can represent a high proportion of the plants in some treatments. What conclusion does the screener draw from a treatment which kills seven out of ten plants of a species in a screen but leaves the remaining three undamaged? Is the chemical inconsistent or is the experimental plant population composed of different biotypes? Or even different species? The wrong interpretation could result in the rejection of a potentially valuable treatment.

SOLUTIONS

In both of the above examples, experimental accuracy, ease of interpretation and confidence in the value of the results would be improved by ensuring, before starting the experiment, that the weed population was clearly defined and appropriate to the objectives of the experiment. How can this best be achieved?

In field trials: Identification of weed seedlings to the species level is difficult with some species, particularly some grasses, Veronica, Chenopodium, Matricaria and Amaranthus. Some practice on live plants, aided by a good key or picture book can overcome problems with these species. A greater problem arises with species in which herbicide tolerant biotypes are widespread. This phenomenon is becoming of increasing importance, particularly in Europe and the USA. A knowledge of the field’s weed flora and agronomy in the previous seasons is a great help in reducing the risks of encountering unwanted population characteristics. A more reliable technique is to select a field in which these species do not occur and to establish the experimental population from seed derived from a correctly identified and well characterised population.

In glasshouse tests: Critical attention to the origin of the seed used to produce plants for glasshouse tests is the key to achieving confidence in the results. The seed must be species-pure and of a uniform biotype appropriate to the objectives of the experiment.

The source of seed: Establishing an experimental population from seed does not achieve a uniform population if the seed itself is not homogenous. Seed collected from farm crop cleanings, or worse, the cleanings from communal silos, will inevitably include several biotypes and possibly several species. It is imperative that the seed used to produce experimental plants is of a single, well characterised population.

THE HERBISEED SOLUTION

Herbiseed employs a well thought out system to ensure that the seed of the major weed species which we supply is well characterised and appropriate to agrochemical experiments. An initial seed stock is collected from an accurately identified and clearly defined population growing in a habitat typical for the species. This is cultivated under controlled conditions to produce the seed which we supply. During growth it is rogued to remove atypical plants. Isolation measures are employed to reduce cross contamination between different populations of the same species. After about five generations, this population is replaced with another population newly selected from the field.

Recently we have installed a database on which the significant characteristics of each seed batch are stored. These can be printed out as a ‘Population Certificate’ to provide the experimenter with essential information to assist in selecting a seed batch and in interpreting the experimental results It also provides documentary authentication of the seed origin for formal GLP/GFP.

APPENDIX1.

1. DIFFERENT POPULATIONS WITHIN A SPECIES

Alopecurus myosuroides. Most field populations in lowland Britain show a level of elevated tolerance to chlortoluron, which varies continuously up to the high tolerance of the Essex 'Peldon' population. Partial tolerance of chlortoluron and some graminicides is also present in some populations in France, Germany and Spain.

Avena fatua. Populations of this species show a large variation in levels of dormancy and seedling frost tolerance. Mother plant nutrition and degree of seed maturity at harvest are also involved in subsequent behaviour of the seed. There is increasing evidence that populations of this species show quantitative and qualitative differences in herbicide metabolism and tolerance.

Galium aparine. A notoriously variable species. 'Northern' and 'Southern' populations differ considerably in competetiveness, habit and reaction to drought. 'Hedgerow' populations have been found to be less tolerant of some herbicides than 'Field' populations. Seed size varies by up to 200% between batches, and appears to have a genetic component. Germination behaviour varies considerably between batches of seed, and that of the same batch varies with time and possibly storage conditions.

Lolium rigidum. Different populations of this species throughout the world exhibit varying levels of tolerance to a range of herbicides. In some cases this is specific to a single herbicide, other populations show tolerance of several unrelated herbicides.

Viola arvensis. There are significantly different populations of this species across Europe, with some indications of genetic interchange with the habit, flower size and colour of V. tricolour. There is significant variation between these populations in their response to hormone and urea herbicides.

Xanthium strumarium. Another species renowned for its variability in morphology, reaction to daylength and germination behaviour. In practice, timing and depth of germination has a significant effect on the ability of individual plants to survive herbicide application. In glasshouse tests, the use of a population which remains vegetative at winter light levels is highly desirable.

2. POPULATIONS OF CLOSELY RELATED SPECIES

A field population of weeds seldom consists of a single species. This can pose difficulties in field trials on natural populations where it will be necessary to identify all of the species present before any treatments are applied. Failure to do so risks missing important information if a treatment is effective against one species but not against a closely similar one.

Seed collected directly from field populations poses similar risks if it is used in glasshouse experiments. Where two or more morphologically similar species grow together, seed collected from that population will not be species-pure, and will give rise to difficulties in interpreting the results of the experiments in which it is used.

Some weed genera in which morphologically similar species growing together give rise to problems of apparantly variable response to herbicide treatments are listed below.

Amaranthus retroflexus. The morphology of this species varies widely across Europe, with spreading branches and compact inflorescences being common in the North and upright plants with lax inflorescences common in the Balkans. Amaranthus chlorostachys is frequently present, unnoticed, in fields infested with A. retroflexus.

In America, mixed infestations of A. retroflexus and A. hybridus occur, with neither species occuring in a pure population except where an atrazine resistant biotype is the only weed present.

Chenopodium album. This species has several sub-species, at least one of which, (C. album amaranticolor) is different to C. album album in its tolerance of certain herbicides. Other superficially similar species such as C. ficifolium frequently grow among populations of C. album.

Matricaria and Anthemis species. These closely related genera tend to be lumped together as 'mayweeds'. However the reactions of individual species to herbicides differ, with A.arvensis and M.recutita being relatively susceptible to hormones, A. arvensis can be more tolerant of chlortoluron and M. perforata is the most tolerant of paraquat. For identification of ‘mayweeds’ see Herbiseed Brief Weed Guide No. 9.

Veronica spp. Many field populations of 'Speedwell' comprise mixtures of V. arvensis, V. agrestis, V. polita and V. persica. These species differ in their sensitivity to herbicides, with for instance, V. arvensis being susceptible to chlortoluron and V. persica being tolerant.

Herbiseed, New Farm, Mire Lane, West End, Twyford, RG10 0NJ, England.
Tel: +44 (0) 1189 349 464  Fax: +44 (0) 1189 241 996  e-mail:[email protected]