Hi All, I find this allometry-ecomorphology topic very interesting and wanted to chime in the philosophical discussion. From my point of view, the main question being asked is how to test the influence of an environmental variable on shape?, given the additional situation that species display significantly different allometric trajectories. In this perspective, phylogenetic relationships and the evolutionary divergence of the 'allometric trajectory' (very well exposed in Esquerre et al.'s article) is less relevant than the actual relationship between environment and shape in the contemporary specimen that you have measured. For instance, I would not find it standardized to use climatic data from, let's say, a thousand years ago to use as a covariate with shape of a contemporary animal, although that might be an appropriate approach if you are using fossil specimens. For this question I think we have to integrate a bit from the field of developmental biology, and ask ourselves when (in ontogeny) does shape react/respond to the environment in a biologically meaningful way. I am not an expert in developmental biology, but I would bet that phenotypic plasticity potential is progressively lost with age. For example, if I feed an adult fish that specializes in suction feeding hard food items, say for several years, it will not change shape and probably die; however, if I rear that same fish from larva, it may develop shape changes that facilitate it feeding on the non-specialized food item. Conceptually, it's as if developmental spandrels appear as age progresses; when a certain anatomical element reaches a certain size, it limits the potential range of changes in interconnected elements. What I'm actually trying to get at here is that even if you use adult shapes, these shapes result from the previous environmental conditions to when you captured the specimens.
So now, what could be a proper methodology to go about the initial question? The most obvious solution is to rear your own ontogenetic series of specimens in controlled conditions; however, we all know that this is almost prohibitive due to space and monetary limitations, and considering the best case scenario of a species with easy access and care. However, I personally don't find it necessary to take into account the allometric-phylogenetic relationship between species to test differences in their response to an environmental variable. Moreover, I believe that the influence of the environmental variable on the phenotype, specially if made to fluctuate beyond natural conditions in the laboratory, will affect the contemporary manifestation of a species' macroevolutionary shape-size relationship. I think that size variation is a non-phylogenetic part of the environment*shape interaction that is inseparable, the interaction effect will depend on what age/size the specimen experienced the measured environmental effect. Nevertheless, the previously known evolutionary relationships may be used to determine what developmental stages are adequate to compare given the presence of heterochronies in a clade. From everything said beforehand, I think that the best way to proceed would be to capture individuals of different developmental stages, trying to cover the largest size range as possible, and during different environmental conditions (probably best to cover the extremes of the environmental variable's range), e.g. different seasons. Later you can do a PLS with the obtained data and see whether the species' response to the environmental variable is different at different stages of its ontogeny (rather than correcting for allometric trajectory); and afterwards compare the same between species. Maybe shape is only responding to the environmental variable meaningfully at a certain time in ontogeny, and then only certain aspects of shape. Apart from the inherent size standardization of the GPA, for these comparisons it maybe interesting to create a new categorical variable for age/developmental stage. If this data is not available, but you have data on the species' growth because of rearing experiments, you could also try to delimit developmental stages using species-specific size ranges. It's also important to point out here that there are different size measurements; and that a large CS does not necessarily correspond (proportionately) to a large length (e.g. SL or SVL). For example, an environmental effect may induce an increase in CS in a species during the juvenile stage that is not accompanied by a proportionate increase in SL or other biometric variable of interest. I also wanted to comment on what Ian said that ontogenetic allometries often have non-linear relationships. I have observed something similar in my own work and wanted to know the opinion of others on the matter. In one of my works I studied the phenotypic development of three fish species, for which I determined (three) developmental stages based on the SL increase in one year observed in rearing experiments. Said differently, I had individuals of all three taxa spanning from 1-12cm SL, at the end of which they had achieved sexual maturity and were adults. When I regress shape variables (e.g. interlandmark distances) on logSL, I observe that these present more of a polynomial relationship within developmental stage groups than the linear relationship you would expect from a typical regression (although I obtain relatively high linear R2 values, they are even larger when adjusting a polynomial). My point here is that the transition between developmental stages is not stepwise. I'm not an expert in math/physics and really don't have the knowledge or vocabulary to go into technical details, but I wanted to ask the community about the possibility of fitting a wave function to the allometric trajectory (or other plots), maybe based on its var-covar structure or other parameter? I have been exploring leisurely for some time now performing ecomorphology on featureless shapes using EFA and may be mixing up theories seriously, but I thought I would ask the mathematical experts in GMM about this crazy idea... I put out a lot of stuff here and hopefully I have explained myself clearly, but I'm sure I'm probably leaving out more than one different perspective and would like to know what others think about the issues raised. Look forward to the responses! Best wishes, Javier ________________________________ From: [email protected] <[email protected]> on behalf of Ian Dworkin <[email protected]> Sent: Wednesday, July 15, 2020 04:43 To: Damien Esquerre <[email protected]> Cc: Mauro Cavalcanti <[email protected]>; Hugo Benítez <[email protected]>; Morphmet <[email protected]> Subject: Re: [MORPHMET2] allometric correction Damien I should also say that given the range of size variation you are suggesting (i.e. where there is no overlap in size distributions between the focal species for instance ) then I would not use the approach I suggested. You would be in a position of extrapolating the shape ~ size relationship likely too far away from where you would be comfortable with the prediction (and uncertainty in the prediction). Also if your intraspecific shape ~ size models are for ontogenetic allometries I am even more skittish (because ontogenetic allometries so often have non-linear relationships). So it may be that what I advocated is not useful for your questions. On Tue, 14 Jul 2020 at 21:32, Damien Esquerre <[email protected]<mailto:[email protected]>> wrote: Dear all, thanks for the excellent feedback. I need to maybe clarify some things: I am definitely aware and agree that interspecies comparisons need to be done in a phylogenetic context. However, when you have multiple individuals per species, and are comparing intraspecific vectors such as allometric slopes between species (I know it sounds contradictory but hopefully you know what I mean) I am not aware of any method that can incorporate phylogenetic information (these usually rely on species means, or one value per species). We sort of explored a bit of that issue in the attached paper. I guess I can think of ways around it, like including 'clade' as a factor in the model, but that doesn't fully account for the relationships between species. I think Dean and Mike have been working on this? I like Ian's suggestion of using species means at a comparable size. However, that wouldn't work when we are talking about species with orders of magnitude in size difference. Something I can think of, even if ontogenetic trajectories are different between species, one could compute the evolutionary allometric trajectory of the mean adult shape of a group of species, and then extract the residuals from that regression. I guess there could be many arguments against doing this if those species have different ontogenetic trajectories, but would love to open a discussion about this. Another option is just not performing any allometric corrections and accept there will be a confounded allometric component to variation with the evolutionary (interspecific variation). In the end, what would you do if you wanted to detect mode of evolution, convergence and or adaptation (effect of environment) in a clade that displays heterogeneous allometric slopes? In particular, when you dont have comprehensive ontogenetic series and that is not the focus of your question. This is something I've been thinking about for years, and it fascinates me, but have never arrived at a satisfactory answer. Thanks again! Damien On Wed, Jul 15, 2020 at 6:04 AM Mauro Cavalcanti <[email protected]<mailto:[email protected]>> wrote: Dear Hugo, >shape of one particular species. Now regarding the Ian example in Drosophila, >I tested allometry into 60 species across the genus and there is definitely a >>pattern (indeed a beautiful one across the genus looking into the different >clades) but looking into one particular group of species the "size" is indeed >a very Have you already published this work? If so, could you please provide a reference? It looks truly great. With best wishes, Em ter., 14 de jul. de 2020 às 16:57, Hugo Benítez <[email protected]<mailto:[email protected]>> escreveu: Fantastic line of discussion I'm absolutely agree with Joe, when there is not a Phylogenetic context maybe the allometric correction would not have very much sense, because we are looking into only one generation so we really don't know very well if the shape we are looking for is product of environmental condition (that can be connected with plenty of variables like nutrition, stress, etc...) now as you asked there is multiple group of species so definitely in your idea there is some "historic factor" that provide the shape of one particular species. Now regarding the Ian example in Drosophila, I tested allometry into 60 species across the genus and there is definitely a pattern (indeed a beautiful one across the genus looking into the different clades) but looking into one particular group of species the "size" is indeed a very good trait to explain differences (like species from island, marsh or the typical cosmopolitan) cosmopolitan Drosophila have the simple small wing (very small). On the other hand, where maybe "makes sense" is in one single species after doing some quantitative genetics experiments and controlling the factors that could influence the size depending on your question... But I think if there is a simple species in the game the factors in one generation are indeed just the real biological meaning of your differences and I dont think a correction will have a very biological meaning... Now of course could be some exception to the rule and a Biogeographical question like bergmann rule, or another rule like that where the relationship is directly related to size maybe a correction could be ok to see how big there are the differences when the factor is included... I would love to see more replies, nice topic to discuss Damien Best Hugo Benítez El mar., 14 jul. 2020 a las 14:55, Damien Esquerre (<[email protected]<mailto:[email protected]>>) escribió: Dear morpho community, I have a philosophical question on size correction that should start an interesting discussion. When we are interested in seeing the effects of species or environmental variables for example, on shape, people often first remove allometric variation by computing the residuals of a shape ` size regression. This of course, doesn't make sense if there are heterogeneous slopes and species have different allometric trajectories (i.e. if the species*size term is significant). What do you think would be the most appropriate way to deal with this situation then, if you are interested in environmental effects on shape? Best regards, Damien Esquerré -- You received this message because you are subscribed to the Google Groups "Morphmet" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]<mailto:[email protected]>. To view this discussion on the web visit https://groups.google.com/d/msgid/morphmet2/CAJiv7Cw0S2A%3D-kpEmknt-tUp92Vjy0hkiF%3DVF6C77NsFECP%2BMw%40mail.gmail.com<https://groups.google.com/d/msgid/morphmet2/CAJiv7Cw0S2A%3D-kpEmknt-tUp92Vjy0hkiF%3DVF6C77NsFECP%2BMw%40mail.gmail.com?utm_medium=email&utm_source=footer>. -- Dr. Hugo A. Benítez Profesor Asociado Centro de Investigación de Estudios Avanzados del Maule Universidad Católica del Maule Research Associate, University of Cambridge Museum of Zoology External Researcher Faculty of Agriculture, University of Zagreb Lab website: http://www.morphoshape.com<http://www.hugoabenitez.com> -- You received this message because you are subscribed to the Google Groups "Morphmet" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]<mailto:[email protected]>. To view this discussion on the web visit https://groups.google.com/d/msgid/morphmet2/CACTC4WqeQiyyLW0gGy%3Ds8jr%3DM0pXAJmuEqXGKUCAaxs3%3DOa%2BmA%40mail.gmail.com<https://groups.google.com/d/msgid/morphmet2/CACTC4WqeQiyyLW0gGy%3Ds8jr%3DM0pXAJmuEqXGKUCAaxs3%3DOa%2BmA%40mail.gmail.com?utm_medium=email&utm_source=footer>. -- Dr. Mauro J. Cavalcanti E-mail: [email protected]<mailto:[email protected]> Web: http://sites.google.com/site/maurobio "Life is complex. It consists of real and imaginary parts." -- You received this message because you are subscribed to the Google Groups "Morphmet" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]<mailto:[email protected]>. To view this discussion on the web visit https://groups.google.com/d/msgid/morphmet2/CAC1JhZYNJdDWqfrsGXGa%2Be0YX8oA5CjoLBTXa67wHuGHaSqJbQ%40mail.gmail.com<https://groups.google.com/d/msgid/morphmet2/CAC1JhZYNJdDWqfrsGXGa%2Be0YX8oA5CjoLBTXa67wHuGHaSqJbQ%40mail.gmail.com?utm_medium=email&utm_source=footer>. -- You received this message because you are subscribed to the Google Groups "Morphmet" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]<mailto:[email protected]>. To view this discussion on the web visit https://groups.google.com/d/msgid/morphmet2/CAJiv7CxV96Zj%2B6Sq793eq%3DB-5EoJ4FbbbOW5K1DsS3ENwYCuDw%40mail.gmail.com<https://groups.google.com/d/msgid/morphmet2/CAJiv7CxV96Zj%2B6Sq793eq%3DB-5EoJ4FbbbOW5K1DsS3ENwYCuDw%40mail.gmail.com?utm_medium=email&utm_source=footer>. -- Ian Dworkin Department of Biology McMaster University Office phone 905 525 9140 ext. 21775 Lab phone 905 525 9140 ext. 20076 [email protected]<mailto:[email protected]> dworkinlab.github.io<http://dworkinlab.github.io> -- You received this message because you are subscribed to the Google Groups "Morphmet" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]<mailto:[email protected]>. To view this discussion on the web visit https://groups.google.com/d/msgid/morphmet2/CAGudrjm_Z6-%3D5Pv7MoFs_JvjWNoS6wPj5Z0PBYViHQPB3fpZmA%40mail.gmail.com<https://groups.google.com/d/msgid/morphmet2/CAGudrjm_Z6-%3D5Pv7MoFs_JvjWNoS6wPj5Z0PBYViHQPB3fpZmA%40mail.gmail.com?utm_medium=email&utm_source=footer>. -- You received this message because you are subscribed to the Google Groups "Morphmet" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/morphmet2/AM6PR03MB55737E6DF5C3C594D7607B8FEB790%40AM6PR03MB5573.eurprd03.prod.outlook.com.
