Compound
 
Interaction (feeding intensity x Herbivore treatment)

Herbivory treatment

Feeding intensity



Variance structure

Likelihood ratio

p. value

Likelihood ratio

p. value

Likelihood ratio

p. value


Monoterpenes
       
αPinene (cyclic)

7

0.906

0.824

15.227

0.002**

15.393

<0.001***

Camphene (cyclic)

2

0.999

0.802

8.246

0.038*

8.883

0.003**

Myrcene (cyclic)

4

0.642

0.887

8.364

0.04*

7.533

0.006**

Borneol (cyclic)

2

0.779

0.855

6.925

0.074

9.052

0.003**

(Z)Ocimene (acyclic)

8

3.951

0.267

7.286

0.063

8.424

0.004**

(E)βOcimene (acyclic)

4

1.384

0.709

9.797

0.020*

14.036

<0.001***

Linalool (acyclic)

2

1.441

0.696

2.456

0.483

11.012

<0.001***

Homoterpene
       
DMNT

8

7.797

0.051

3.786

0.286

0.444

0.505

Sesquiterpenes
       
(E)βCaryophyllene

2

0.667

0.881

11.371

0.01*

4.334

0.037*

αHumulene

2

1.526

0.676

9.014

0.029*

1.676

0.196

Nerolidol

4

6.410

0.093

10.387

0.016*

12.891

<0.001***

Green leaf volatiles
       
(Z)3Hexenyl acetate

2

0.656

0.884

6.067

0.108

6.454

0.011*

(Z)3Hexenol

1

2.284

0.516

16.015

0.001**

0.138

0.71

Ncontaining comp.
       
2Methylbutyraldoxime

7

0.522

0.914

10.821

0.013*

4.454

0.035*

3Methylbutyraldoxime

4

0.536

0.911

14.950

0.002**

10.335

0.001**

Benzyl cyanide

3

2.723

0.466

10.852

0.013*

9.464

0.002**

Indol

3

2.136

0.545

9.688

0.021*

1.537

0.215

Aromatic compounds
       
Salicyl aldehyde

8

8.734

0.033*

13.706

0.003**

0.535

0.464

Benzyl alcohol

4

4.867

0.182

4.624

0.202

6.770

0.386

Benzene ethanol

4

8.003

0.046*

7.629

0.054

0.703

0.402

 For each parameter the F and p values are given. Asterisks indicate significant differences, p < 0.001 = ***, p < 0.01 = **, p < 0.05 = *, no asterisk = not significant. Compounds are grouped according to their chemical classes, the second column shows the variance structure with the lowest Akaike Information Criterion (AIC), which was used in the Generalized Least Square model (GLS). Variance structures tested were as follows: 1. varFixed variance for feeding intensity, 2. varIdent variance for herbivory treatment, 3. varPower variance for herbivory treatment, 4. varExp variance for feeding intensity, 5. varConstPower for feeding intensity, 6. varConstPower for feeding intensity and herbivory treatment, 7. Combined variance (varIdent for herbivory treatment, varFixed for feeding intensity) and 8. Combined variance (varIdent for herbivory treatment, varExp for feeding intensity). A detailed description of the variance structures is given by [80].