Trial layout & design

physical layout of a trial sections trials are commonly set out in sections, which are contiguous physical areas on a trial site these are typically identified on the ground by numbered pegs for each tree, gem records its section and its position within that section the latter is expressed as an integer which is normally a relative position within a systematic numbering convention it is recommended that documentation of the numbering convention used should be stored in the system and linked to all trials that use that convention experimental design of a trial based on genetic treatments trial entries the same genetic entity can have different roles in different trials (e g , be a entity under test in one trial and a control in another); or be produced in different ways (e g , by different orchards, by different vegetative propagation methods, etc) in addition, the same genotype can be tested as a replicated copy of itself (a clone) or as a parent of seed progeny in all of these cases the same entity is being tested the participation of a genetic entity in one trial in one experimental design role is referred to as a trial entry the same genetic entity can have entries in multiple trials, and may even have multiple entries in the same trial if the experimental design requires it to be tested in more than one role the following parameters collectively characterise an individual trial entry, in other words they define the role under which the genetic entity participates in the trial field codes each trial entry must have a field code (an integer, unique within the trial but not necessarily beginning from 1) field codes are used as convenient shorthand references to a trial entry the same genetic entity will usually, but not necessarily, be assigned different field codes when it participates in different trials experimental role each trial entry has an experimental role experimental roles are defined as entry under test (eut) genetic entities, that is, materials originating from selections, full sib families, and seedlots, can be defined as eut if the aim of the trial is to compare their performance control (ctl) a control is usually a genetic entity with known performance, which is put into a trial for comparison connector (con) some testing programmes have identified one or more groups of genetic entities that are included in every trial in order to ensure connectedness any or all of these roles may be used in any trial, though use of the connector role is expected to be very limited experimental roles may be outputted in data extracts and can, therefore, be used to subset data if desired also, system generated summary statistics are calculated for each experimental role separately as well as in all combinations tested as each trial entry has a classification called tested as this variable distinguishes the testing of vegetatively propagated clones from the testing of a group of genotypes which have one or more parent in common allowed values for tested as are c (clone) the entity consists of cloned ramets of a genotype from the selection register h (half sib) the entity is a group of progeny all of which have one parent in the selection register in common this would include open pollinated seed from a mother in the selection register, and seed from controlled pollination of a single parent with a pollen mix f (full sib) the entity is a group of progeny from a full sib family listed in the full sib family register, which in turn identifies maternal and paternal genotypes in the selection register s (seedlot) the entity originates from a seedlot in the seedlot register for some purposes, such as certain pedigree extracts , it is important to identify individual genotypes explicitly when reporting on the trees associated with a trial entry the unique identifier is referred as a genotype code propagation methods trial entries may also be classified by propagation method propagation methods are defined by the user examples might be seed, stoolbed cutting, embryogenesis, bare root, container, or any other definition provided by the user other experimental design parameters gem can accommodate complex experimental designs, with or without a genetic component if there is no genetic component, there may be only one trial entry, associated with a single genetic entity (which may even be an 'unknown' entity) experimental design parameters that are not associated with a trial entry are specified by user defined fields (udf) these are optional classifications that are specific to a trial, and can be specified at the section level and/or at the individual tree level udfs are outputted if required in an extract file up to eight udfs can be defined at the section level each can be used as a separate experimental design parameter (e g replication, incomplete block, treatment, split plot, fertiliser treatment, silvicultural treatment, etc ) with a value for each assigned to each section in addition, another eight udfs can be assigned to each individual tree in a trial these can be used as additional experimental design parameters (e g tree position parameters such as rows and columns) the flexible combination of experimental design characteristics of trial entries, plus udfs which are statistically independent of the trial entries, allows enough flexibility for specification of very complex experimental designs