Hello, Liam.

I have a similar problem in a trial to asses the biomass-for-energy production of woody crops installed at high density.

I do not have an answer to your question, because I did not yet found one for myself and my objectives are different from yours. My option was to correct the final density of plants for mortality along one year after establishment (i. e., the final density is different from the initial), but that was because I could do it from an experimental design point of view: mortality is one of the parameters to evaluate (the other is biomass increase in an area basis).

Therefore, I did not replace the seedlings to avoid that bias, but because I am comparing rain fed and summer irrigated treatments (in Portugal summer is hot and considerably dry), I am now facing the need to replace some of the rain fed treatments, at least for the species most affected by summer drought stress.

The only thing I remember as a suggestion is to consider the 3/2 competition law and use some type of statistical comparison between the respective regressions, but I am not an expert in statistics and I am not sure thar this will work properly...

Dear Liam,

I am not sure I understand what you mean by " will be deployed mono clonally and in a mixed mosaic of single tree plots" - are you planning to mix clones on a site, but treat each individual as a single repetition? If so, "natural" mortality is one (extreme) form of competitive strength (interclonal competition) and therefore there is - in my opinion - neither need nor want for any replacements. The clone does not grow in your trial - why should it grow in "real life"? If you want to compare growth on an inter-clonal basis, the best way to go is (again: in my opinion) to have areas with pure stands (mono clonal, but not single tree) and areas with mixed stands (multi clonal, multi tree). Thus, you can compare the effects of inter- and intra-clonal competition.

Best regards,

Joachim

HI António / Joachim.

My design is to create monoclonal stands, each will be comprised of one family and then to compared the competitive ability of each family, they will be mixed in single tree plots, which

, sorry for not being clear, will be a mixed stand of all families, The idea would be to try and identify crop and competition ideotypes as well as comparing their performance and resource allocation. For a balanced analysis I would require all trees to remain alive to gauge their response, if for example all of the clones from one family died from say transplant shock then I would not be able to make a comparison ( other than say one died, I think). This would increase the growing space of the neighbouring trees not due to their competitive ability but only cause of their location. My main concern is that during the establishment phase mortality for one family maybe high and thus it may not become established in sufficient numbers and I am wondering if this is the case and ramets were left without competition from one side for say four weeks, they would have a competitive bias compared to others who had competition during this time. Four weeks out of one growing season may give them a sufficient advantage especially considering the trial is planned to last only 3 growing season (c. 1/36th).

I hope this is clearer.

Thank you both for your time and suggestions,

Liam

Hello Liam,

unfortunately, I am still at a loss as to what is the hierarchy of your experiment - but I am not a geneticist. As far as I got it, the single tree which is planted *somewhere* belongs to the same species, but to a different clone and family? And each family is also planted in a pure stand with n individuals? If this is the case, you could compare mortality of a clone in the pure and in the mixed stand. Soil and weather being equal, any difference in mortality is likely due to competition or treatment.

For the mixed stand, I would chose a triangular planting scheme. This way, each single tree has six neighbors and each neighbor potentially has the same space to exert competive pressure. If some fail you can include this in the analysis.

Best regards,

Joachim

Dear Liam

I am not clear about your question

Normally if you are assessing the clonal performance the mortality is one of the important factor of assessing its performance.

In your question you have not mentioned anything about the purpose of your trial

I do have have considerable experience in clonal trial I can provide some useful tips if I get the details of your experiment along with objectives and treatment/clonal details

thanks

we measure the total biomass from a plot. In this way, we deal with the compensatory growth that arises when there is mortality in a plot. The one dead tree is not "lost" completely because of the compensatory growth of the neighbours, but it does result in an overall loss of biomass from the plot. If you try and measure mass per tree, then you end up with errors because of compensatory growth, making some individual trees bigger than trees in a plot with no mortality.

Hi Liam,

I am not familiar with this kind of trials, but maybe you could think about planting seedlings in some regular way in experimental plots but, at the same time, to plant some extra seedling in vicinity (same conditions) that could replace dead seedlings in experimental plots. In that way you could have mortality data and protect regular density across plots. Of course, these substituted plants would not be in analysis (something like missing values).

Hi Liam,

I was going to suggest exactly what Stevan above has suggested. What I would do is take survival rate as your first result. Then fill gaps with standby replacements, to keep competitive pressure on neighbours. Clearly label these as replacements and exclude from analysis. SAS analysis of sinlge tree plots can accomodate different numbers of observations.

Dear Liam

I was going through the answers from different researchers and it seems that everyone contributed considerably. By this time you must have cleared your doubt

Otherwise I would suggest the following

Based on your question and clarification I conclude that you are trying to assess the competitive ability of different clones based on their performance in block planting and when they are mixed with other clones .

1. When the clones are in blocks you are going to average the performance of all the individuals present and which will be compared with other clones

2. When the clones are mixed with other clones you have to go for single tree replication and you should go for as many as replication preferably more than 15

Based on your observation you can go for statistical analysis either in Genstat or SAS statistical software which has provision for single tree plot analysis