Expl. Variable
|
Molecular Form
|
---|
M
|
S
|
---|
logit
|
probit
|
log-log
|
logit
|
probit
|
log-log
|
---|
CONC.
|
60.26
|
59.23
|
59.24
|
55.71
|
55.59
|
107.29
|
log(CONC.)
|
141.41
|
166.22
|
108.86
|
27.42
|
27.73
|
33.62
|
- All models have a binomial errors structure but differ in the scale of the explanatory variable (CONC., i.e. ammonia concentration, linear vs. log-transformed), or the link function of the linear predictor (logit vs. probit vs. complementary log-log). The minimal adequate model for each taxon, i.e. that having the lowest AIC, is shown in bold.