326
unpaired, and not encapsulated swim bladders with a relative swim bladder length
(calculated according to the formula: bladder length + height + width/3) of 8.4–15.2 %
of the fi shes’ standard length (Fig. 3a ). Members of these families typically have four
Weberian ossicles (with exceptions) and an ossicular chain length comprising 3.5–5
% of their standard length (Fig. 4a ) (Lechner and Ladich 2008 ). In contrast, representatives of the families Loricariidae and Callichthyidae have signifi cantly smaller swim
bladders (relative bladder length: 1.6–5.7 % of standard length), just one to two auditory ossicles, and thus a signifi cantly shorter ossicular chain (relative chain length:
1.1–1.9 %) (Figs. 3a and 4a ).
How do differences in swim bladder and ossicle structure affect hearing in catfi shes? Measurements of auditory sensitivities using the auditory evoked potential
(AEP) recording technique revealed that all 11 species investigated detect tone
bursts between 50 Hz and 5 kHz (Fig. 5a ). The lowest absolute auditory threshold
was found in the ariid catfi sh (67 dB re 1 μPa), and the highest in a callichthyid (121
dB re 1 μPa). Mean auditory thresholds of all six species having large bladders and
of all fi ve species having tiny paired bladders revealed signifi cant differences in
hearing sensitivity between both groups between 1 and 5 kHz but not at lower frequencies (Fig. 5b ).
Furthermore, the relative swim bladder length was negatively correlated to the
hearing threshold at frequencies above 1 kHz ( r -values between −0.71 and −0.80).
Species with larger bladders were more sensitive than others. Similarly, a longer
ossicular chain and a higher number of ossicles resulted in better hearing at 3–5 kHz
( r -values between −0.61 and −0.67) (Lechner and Ladich 2008 ). These data indicate
that larger, free swim bladders, and longer ossicular chains affect hearing in catfi shes positively by increasing the hearing sensitivities at higher frequencies.
3 Thorny Catfi shes
The large Amazonian catfi sh family Doradidae or thorny catfi shes (~150 species)
exhibit a large intrafamilial variation in swim bladder morphology. This organ is
always unpaired, of different size, and may bear a caudal sac (secondary bladder)
and numerous partly branched diverticula (Birindelli and Sousa 2009 ; Kaatz and
Stewart 2012 ). Doradids are a potential model to investigate the role of the diversity
in swim bladder morphology on hearing within one family because, in contrast to
non-otophysines such as holocentrids, sciaenids, or cichlids, the factor distance
between the swim bladder and the ear is negligible. Their bladders are always
directly connected to the inner ears via a chain of ossicles. Morphological measurements of the swim bladders revealed that species can be grouped according to relative swim bladder size into two groups, namely those species having smaller and
those having larger bladders (Zebedin and Ladich 2013 ) (Fig. 6 ).
The intrafamilial differences in relative swim bladder length in thorny catfi shes
are, however, much smaller than the interfamilial difference among catfi shes
described above. Zebedin and Ladich ( 2013 ) found that relative bladder length in
F. Ladich
unpaired, and not encapsulated swim bladders with a relative swim bladder length
(calculated according to the formula: bladder length + height + width/3) of 8.4–15.2 %
of the fi shes’ standard length (Fig. 3a ). Members of these families typically have four
Weberian ossicles (with exceptions) and an ossicular chain length comprising 3.5–5
% of their standard length (Fig. 4a ) (Lechner and Ladich 2008 ). In contrast, representatives of the families Loricariidae and Callichthyidae have signifi cantly smaller swim
bladders (relative bladder length: 1.6–5.7 % of standard length), just one to two auditory ossicles, and thus a signifi cantly shorter ossicular chain (relative chain length:
1.1–1.9 %) (Figs. 3a and 4a ).
How do differences in swim bladder and ossicle structure affect hearing in catfi shes? Measurements of auditory sensitivities using the auditory evoked potential
(AEP) recording technique revealed that all 11 species investigated detect tone
bursts between 50 Hz and 5 kHz (Fig. 5a ). The lowest absolute auditory threshold
was found in the ariid catfi sh (67 dB re 1 μPa), and the highest in a callichthyid (121
dB re 1 μPa). Mean auditory thresholds of all six species having large bladders and
of all fi ve species having tiny paired bladders revealed signifi cant differences in
hearing sensitivity between both groups between 1 and 5 kHz but not at lower frequencies (Fig. 5b ).
Furthermore, the relative swim bladder length was negatively correlated to the
hearing threshold at frequencies above 1 kHz ( r -values between −0.71 and −0.80).
Species with larger bladders were more sensitive than others. Similarly, a longer
ossicular chain and a higher number of ossicles resulted in better hearing at 3–5 kHz
( r -values between −0.61 and −0.67) (Lechner and Ladich 2008 ). These data indicate
that larger, free swim bladders, and longer ossicular chains affect hearing in catfi shes positively by increasing the hearing sensitivities at higher frequencies.
3 Thorny Catfi shes
The large Amazonian catfi sh family Doradidae or thorny catfi shes (~150 species)
exhibit a large intrafamilial variation in swim bladder morphology. This organ is
always unpaired, of different size, and may bear a caudal sac (secondary bladder)
and numerous partly branched diverticula (Birindelli and Sousa 2009 ; Kaatz and
Stewart 2012 ). Doradids are a potential model to investigate the role of the diversity
in swim bladder morphology on hearing within one family because, in contrast to
non-otophysines such as holocentrids, sciaenids, or cichlids, the factor distance
between the swim bladder and the ear is negligible. Their bladders are always
directly connected to the inner ears via a chain of ossicles. Morphological measurements of the swim bladders revealed that species can be grouped according to relative swim bladder size into two groups, namely those species having smaller and
those having larger bladders (Zebedin and Ladich 2013 ) (Fig. 6 ).
The intrafamilial differences in relative swim bladder length in thorny catfi shes
are, however, much smaller than the interfamilial difference among catfi shes
described above. Zebedin and Ladich ( 2013 ) found that relative bladder length in
F. Ladich