wilcox.test usage in R I would like to know if there is a difference in the values taken from measurements of a given characteristic sampled at primary forest (Point=1) and secondary forest (Point=0)
I'm using two vectors, one with the data itself (DATA$Measurements), and the other one with the category of sampling point (DATA$Point).
How should I perform the wilcox.test fucntion?
I'm doing this way:
wilcox.test(DATA$Measurements~DATA$Point)
But I notice that are other forms of performing this test:
wilcox.test(DATA$Measurements,DATA$Point)
wilcox.test(DATA$Measurements,DATA$Point,paired=TRUE)
Which one is more suited for testing my question of interest (Is there a difference between points for this given characteristic), and what is the difference between these options??
 A: This is a question of whether your samples are paired or independent.  Since one sample comes from a primary forest, and another from a secondary forest, it sounds like the samples are independent.  This would be true if the two samples are collected in such a way that there is no logical connection that links an individual from the primary forest to a unique individual from the secondary forest.
However, sometimes researchers create artificial pairings based on similar characteristics of two sites.  For example, suppose a primary forest site (say site A1) is remarkably similar to a secondary forest site (say site B1) based on other measured characteristics, and so they are paired together.  If this is done systematically for the entire sample, then you could want to do a paired sample test.  (Please forgive my lack of field specific forest knowledge)
If the former description of your sample is true, you would use the "paired = FALSE" option.  The latter description would be a situation where "paired = TRUE" would be used.

Regarding the syntax:
wilcox.test(DATA$Measurements~DATA$Point)

The above should be used in your case, since you want to test measurements by "Point" classification.  This is called using the formula interface.
wilcox.test(DATA$Measurements,DATA$Point)

The above tests if there is a difference between two explicitly stated vectors of data (specifically DATA\$Measurements compared with DATA\$Point).  This test would be nonsensical.
You could do 
wilcox.test(DATA$Measurements[DATA$Point == 1],DATA$Measurements[DATA$Point == 0])

to test the difference in Measurements where Point = 1 to those where Point = 0.  This would be equivalent to using the formula interface.
