I have done tracksphere experiments of a pest caterpillar with Syntech Tracksphere LC 300. I have recorded the responses of the caterpillar towards 6 different plant leaf extracts, extracted in dichloromethane (uninfested and infested leaf extracts of 3 varieties). Five replicates were recorded for each extract, each time with a new caterpillar (separate recordings were done for dcm alone and for air). Can anyone please suggest a method for its statistical analysis?
Dear Shameer,
it is a long time ago that I have worked with the Syntech Tracksphere LC 300 (Geiselhardt et al 2008a/b), but I think the statistics have not changed that much. I have used the modified Rayleigh V test for directional data, but see references for more details. I can highly recommend the listed books to become acquainted with circular statistics. In the meantime, there are R scripts available to do the circular statistics (see web link). Anyhow, I am afraid that your sample size is to small to get significant results, unless you have very clear results.
Best wishes,
Sven
Geiselhardt et al. (2008) A sex pheromone in the desert tenebrionid beetle
Parastizopus armaticeps. J Chem Ecol 34
Geiselhardt et al. (2008) 1-Tridecene—male-produced sex pheromone
of the tenebrionid beetle Parastizopus transgariepinus. Naturwissenschaften 95
Batschelet (1981) Circular Statistics in Biology. Academic, London.
Fischer (1993) Statistical anaylsis of circular data. Cambridge University Press.
Pewsey et al (2013) Circular statistics in R (http://circstatinr.st-andrews.ac.uk/)
Dear Dr. Sven Geiselhardt,
Thank you so much for your very valuable suggestions. I will certainly look up the references you have given. I am extremely sorry for the delay in sending my reply, because I was on vacation for 10 days.
Thanks once again.
Shameer K.S.
University of Calicut
Because it is possible to measure many parameters from the 2D tracks that are recorded with Syntech tracksphere there is not a single response.The first point to consider is what kind of data you will extract from the raw tracks.
You might be more interested in directionality, final vector length and vector angle, or mean angles that require circular statitstics. But in case you need to compare two treatments that appears to show some attractivity, other parameters that measure a response "intensity" may be calculated directly or with R scripts working from the csv files: total walked distance, speed, both can be treated by standard statistics. And you can also make distribution stats on the numbers of animals that for instance walk a minimal distance or show positive attraction.
I agree with Sven that 5 replicates is far from enough. Due to variability in insect behaviour, 30 replicates per treatment are recommandable. You can find examples of application in our paper Party et al 2013 PlosOne (http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0052897)
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publication