Flavour quality (Table 4) was influenced by breed
(P 6 0.01), but less than for tenderness for the SM muscle.
F-values among muscles were very similar, although
significance levels differed. As shown in Table 5, light
and early fattening lambs of the Churra breed aged
for two or more days had lower flavour quality in ST
and GB muscles but suckling lambs obtained the best
scores for SM, although the differences were not always
significant. This could be because Churra is a dairy purpose
breed. It has a faster development, so it may develop
more strong mutton flavours at higher slaughter live
weights, which are undesirable for most consumers in
the Mediterranean area who prefer milder flavours.
Slaughter live weight (Table 4) continued to have, in
general, the most important effect on flavour quality,
but much less than for tenderness. Thus, suckling lambs
(Table 5) were considered to have a more desirable flavour
in more than 65% of the cases (9, 10 or 11 cases
out of 15 – three breeds and five ageing times – for
ST, SM and GB, respectively), but differences were not
always significant, according to Duncans test. In a survey
of Canadian consumers, Jeremiah, Tong, and Gibson
(1998b) found that the mid-weight group of their
study (40.5–67.7 kg) had the highest flavour scores,
although differences in slaughter weight percentage were
lower with respect to our work.
Ageing (Table 4) had little effect on the flavour quality
of SM and ST, but an important effect on GB
(P 6 0.001). Spanish Merino GB flavour (Table 5) improved
at longer ageing times, developing desired flavour
precursors. The methodology followed in this
work seemed to be effective in avoiding bacterial con-tamination and spoilage, which could reduce flavour
quality due to undesirable flavours (Reineccius, 1979).
As with tenderness, consumers gave the best scores
for flavour quality (Table 5) to the ST muscle aged for
1 or 2 days. The GB muscle tended to have a more desirable
flavour at longer times of ageing, as mentioned
above. SM also got the lowest scores for this attribute.
We did not observe mean scores that were not acceptable
(5.0 and below).