Maturity acceleration by S. carnosus…
varians (LM1) was used as starter culture; in the experimental low, red, blue, and green. NCS color notations are based on how
batches, a mixture of L. curvatus and S. carnosus 833 was used. In closely the color of the tested sample resembles the 6 elementary
both cases the mixtures of starter cultures were added at a dose colors (Scandinavian Color Institute 2001).
of 0.05% w/w in order to have 1010 bacteria/kg mince. The meat
Before the sensory evaluation, the sausages were defrosted for
and fat cuts were chopped in a cutter into 2- to 3-cm pieces, 2 d at 5 ЊC. Samples were given 3-digit randomized numbers for
minced to 3.5 mm, and mixed with the other ingredients and ad- coding, and served as 0.5-cm-thick pieces at room temperature
ditives. The freeze-dried starter cultures suspended in a mini- in random order according to sample and assessor. All samples
mum amount of water were added to the minces, and the minces were analyzed twice by the sensory panel. Each assessor evaluat-
were stuffed in reconstructed collagen casings of 6.5-cm dia in ed the samples at individual speed on a computerized system for
650-g pieces. Fermentation was carried out at 21 to 22 ЊC, 95% direct recording of data (CSA Compusense ver. 5.38, Canada).
relative humidity (RH) for 12 h, and was followed by drying for 4 The computer transformed the responses into numbers from 1.0
d at temperatures gradually decreasing from 21 to 22 ЊC to 17 to = no intensity to 9.0 = high intensity. The samples were analyzed
18 ЊC and at RHs decreasing from 85 to 90% to 70 to 80%. Final for the following attributes: overall intensity of odor, odor of pork
maturation took place at 12 to 13 ЊC and at 83 to 87% RH. pH of meat, sour-sweet (sour, fruity-sweet) odor, odor of sour socks,
the finished sausages was measured on a 2 g sample homoge- odor of spices, odor of pepper, odor of garlic, maturity odor,
nized in distilled water (water/sample, 10:1) with a Xerolyt elec- whiteness, hue, color intensity, acid taste (basic taste sour), salty
trode (52-32, Crison Instruments S.A., Barcelona, Spain) connect- taste, sweet taste, bitter taste, metallic taste, overall intensity of
ed to a Ph meter (pHM82, Radiometer, Denmark).
flavor, meat flavor (pork), sour-sweet (sour, fruity-sweet) flavor,
spicy flavor, pepper flavor, garlic flavor, maturity flavor, rancid
flavor, flowery flavor, nutty flavor, cheesy flavor, hardness, ten-
derness, juiciness, fattiness, stickiness, granularity. Twelve sau-
sages were evaluated, 3 of each type: standard sausages—STD-
40, STD-55, and STD-70 (ripened 40, 55, and 70 d), and
experimental sausages—EXP-40 (ripened 40 d).
Volatile analysis
Two sausages from each of the batches matured for 40 d were
analyzed twice by triplicate analysis in randomized order. The
sausage samples were cut into smaller pieces and 80 g of sausage
was minced with 55 g of water in a small domestic food proces-
sor (Krups GmbH, Solingen, Germany) for 1 min. Twenty grams
of sausage slurry plus 80 g of water were weighed into a 125 mL
glass bottle with a purge head. The slurry was equilibrated for 30
Statistical analyses
Differences between sausage types regarding volatile content
min in a water bath at 42 ЊC, purged with N2 (50 mL/min, purity and sensory scores were analyzed by analysis of variance (ANO-
99.999%, AGA, Ballerup, Denmark) for 30 min at 42 ЊC, and the VA) and by principal component analysis (PCA) – the PCA in Un-
volatiles trapped onto 225 mg of Tenax TA, 60-80 mesh scrambler, version 7.5 (CAMO ASA, Trondheim, Norway). For the
(Crompack, Knebel, Denmark). Triplicate Tenax TA tubes were PCA, the volatile matrix consisted of the relative peak areas in
made for each sample. Volatiles on Tenax tubes were desorbed in the logarithmic10 scale. The sensory and volatile profiles were
a 2-step manner by an automatic thermal desorber (ATD400, correlated by partial least squares regression (PLSR), and the sig-
Perkin-Elmer Ltd., Beaconsfield, England) and injected directly nificant volatile compounds and descriptors picked out by jack-
onto a DB-1701 column (0.25 mm ϫ 30 m, 1-mm phase, J&W Sci., knifing on a p Ն 0.05 significance level (Unscrambler version 7.5,
Köln, Germany) in a GC-MS system (HP 5890 Series II–HP 5972, CAMO ASA, Trondheim, Norway). Jackknifing is a significance
Agilent, Palo Alto, Calif., U.S.A.). Desorption temperatures were test where t-testing is performed for each PLSR regression coeffi-
tube, 200 ЊC; cold trap, 250 ЊC. The transfer line temperature cient relative to the square root of its estimated uncertainty vari-
from ATD to GC was 200 ЊC. The temperature program in the GC ance. On the basis of jackknifing, unreliable volatile or sensory
was 35 ЊC for 1 min; 4 ЊC/min to 175 ЊC, 10 ЊC/min to 250 ЊC, and variables were eliminated to simplify the final model and making
250 ЊC for 5 min. MS parameters were ionization energy, 70eV; it more reliable (Martens and Martens 2001).
scan range, 33 to 300 AMU; transfer line temp., 280 ЊC. Identifica-
Results and Discussion
tion was based on MS spectra compared with the NBS75k data-
base (Hewlett Packard, Agilent, Palo Alto, Calif., U.S.A.) and on
Kovats retention indices of authentic compounds. The desorp-
tion procedure was calibrated at every session by running three
calibration tubes containing 5 mL of a 0.1 g/L octane in methanol
solution. The calibration tubes were prepared according to Per-
kin-Elmer (1991). Peak areas from the experimental tubes were
divided with the averaged area of the octane peaks from the 3
calibration tubes and multiplied by 106 to obtain a relative peak
unit of 0.5 g octane equivalents.
Sensory profile
Table 1 shows the results from the sensory analysis of the
standard sausages matured for 40, 55, or 70 d (STD-40, STD-55,
STD-70, see “Experimental design” section) and the experimen-
tal sausages with S. carnosus matured for 40 d (EXP-40). Accord-
ing to the analysis of variance, the 4 sausage types were in gener-
al significantly different regarding most descriptors in
appearance, texture, taste, odor, and flavor. Only overall odor
and flavor intensity, sweetness, peppery odor and flavor, garlic
odor, and flowery and nutty flavor seemed to be the same for all
samples.
Figure 1 shows the bi-plot from a principal component analy-
sis on the sensory profile excluding the nonsignificant descrip-
tors. The total plot explains 94% of the variance within the senso-
ry data, but 92% of the total variance is explained by the first axis
due to the great differences between sausages matured for differ-
ent periods of time. The number of maturing days increases from
left to right for the standard sausages STD and divides the sau-
sages into 4 separate clusters—though the distance between the
Sensory analysis
Sensory analysis was carried out in facilities of standards re-
quired in ISO (1988) and ISO/IEC (1990). A standard flavor profil-
ing method (ISO 1985) was used for testing the sausage samples.
Ten assessors, selected and trained according to ISO (1993), fol-
lowed the procedures for quantitative descriptive testing. The
panel was trained in the use of the glossary necessary for the spe-
cific task. The color attributes were evaluated following the Natu-
ral Color System (NCS). NCS is a color system building on human
perception of color. The 6 elementary colors are white, black, yel-
Vol. 67, Nr. 5, 2002—JOURNAL OF FOOD SCIENCE 1915