Statistical methods of detection of a periodic phenomenon in a short series. Example of application: a density series of nematode eggs - II.

Using the example of the statistical treatment of a series of density of nematode eggs, we give here a survey of the main methods of detection of a periodic phenomenon in a short series. The purpose of these methods, which may be ranked into 3 groups: smoothing methods, regression methods, autocorrelations, is to study non random characteristics of the process. We detected in this series and estimated a rhythm of period 12 h. We concluded that the regression methods are the most resourceful for detecting a deterministic periodic phenomenon but that it is useful to confirm the results using the other methods.