170
C. Zimmennann and G. Hubold
A different approach to the data analysis is shown in Fig. 3: instead of
factor loadings, the component scores are now projected into the factor
planes. Component scores are the coordinates of the 20 single experiments
(or fish individuals) in the factorial space. The axes correspond to those in
Fig. 2. The ordination of the single experiments permits the ecological
interpretation of the data.
200
•
ISO
;
• 1.00
1;
1 050
~
• f,I 000
! .(ISO
u
•
II. · 100
· 1 SO
·2.00
a
. 9oteoAa
MyoxoA •
• BoreoAc
• MyoxoD
e
e BoreoAb
PogoO
Myoxoe
e PogoA .
e Myoxoa e AnarhA
e PogoB ~ogoC
e BotchG
e 90tchE
e BotchF
BemacA e BorchC e
BotchO
e GymnoA
· 2.50 ·2.00 · 1.50 . , 00 .(I.SO 0.00 050 1.00 1.50 2.00
"
~
0
- U
II.
e PoooB
·o.SO
. , 00
. , SO
. PogoA
F.ctor 11 nActlvlty"
e pogoC
. BotchG
BorchF
BordlA .
BoreoAb
e e e BoreoAc
GymnoA MyoxoAe
e BorchO
e BotchC
MyoxOO e, AnarhA
yoxoe
·200 +-~r-~--~------r--'--~------+
·2.SO ·2.00 . ,.50 · 1.00 ·0 SO 0.00 0 SO '00 1.50 200
Fig. 3. Factor analysis: plot of the experiment's component scores in factorial planes 1·2
(a) and 1-3 (b). The grey outlined arrow represents a suggested "mode of life" gradient.
Factors were named according to the characterizing variables. Abbreviations of are the
same as in Table 1