Project report Sensory qulity nd shelf life of trditionl utters compred to high qulity industrilly produced utters Agnieszk Dudkiewicz* (University of Lincoln, Ntionl Centre for Food Mnufcturing, Prk Rod, PE 7PT, Holech, UK) *corresponding uthor (dudkiewicz@lincoln.c.uk)
. Introduction In UK s consumer perception, food sold loclly y smll mnufcturers e.g. frmer s mrkets is fresher, tstier nd might e of higher qulity thn the produce retiled in supermrkets (McEchern et l., ; Spiller, ). Frmers mrkets offer rnge of rtisn nd trditionl foods tht re uncommon in supermrkets nd these products re perceived y consumers s mde with more cre nd thus re highly vlued (Autio et l., 3). Despite modern consumer s interest in frmer s mrkets, trditionl nd loclly sourced foods informtion on the qulity nd shelf-life of such produce is very limited. An exmple of such food product is trditionl utter. Trditionl utter is perceived s luxury product nd seems to e vlued for specil sensory ttriutes y consumers nd resturnts (Higgins nd Ptterson, 3). Although there is no informtion ville on trditionl utter consumption or sles in UK, online serch reveled tht there re severl frm shops trding the product tht cn e found online. The we descriptions of the trditionl utter indicte tht it is product tht is mde y frmer from the milk of their own cows, often y hnd. However, it is not well understood how this trditionl process differs from lrge scle industril process. From previous short consultncy reserch work tht ws conducted t Ntionl Centre for Food Mnufcturing we hve found tht smple of trditionl utter crried more thn. totl microorgnism count in g of product. Typiclly the utter should crry only up to microorgnisms in g (Adms nd Moss, 8), lthough this might vry for utters mde from cultured crem (Neves nd Lngridge, 998). Despite n uncertin reson for this prticulr smple crrying high microiologicl count, it cn e speculted tht the incresed numer of production steps nd prolonged opertion times required to mke utter y hnd my e contriuting fctor. Some uthors lso mention tht in very trditionl utter production where the wooden churners could e used, microiologicl contmintion of utter might e incresed. This is due to difficulty of snitising wood compred to stinless still or luminium churners (Budkhr et l., ). The introduction of microiologicl contmintion during utter mnufcture is likely to reduce its shelf life. Butter generlly crries low microiologicl lod s most of the wter contining microes is seprted out in the production process (Adms nd Moss, 8) Furthermore utter provides only very low concentrtions of nutrients for microorgnisms, limiting growth of mny (Adms nd Moss, 8). Despite this, microiologicl spoilge of utter my hve plce due to contmintion with cteri of genus Pseudomons spp. s well s some yests nd moulds tht re le to survive nd grow t low tempertures (Budkhr et l., ). The min reson for the termintion of the utter s shelf life is development of the ft rncidity. Ft rncidity progresses with the utter storge nd is result of oxidtion s well s lipolytic ctivities of microiologicl nd originting from milk enzymes (Munro et l., 998). Level of the ft degrdtion is relted to utter s sensory qulity (Munro et l., 998; Chmpgne et l., 99) nd thus requires monitoring in studies on qulity nd shelf-life of utter. The im of this study ws to conduct sensory, chemicl nd microiologicl nlysis on trditionl utters nd compre it to high qulity industrilly produced utter in order to nswer three questions:. Is the sensory qulity of trditionl utters superior to industril utters?,. Does trditionl utter spoil fster thn industril utter? 3. Wht re the resons for ny differences etween trditionl nd industril utters?.
This study ws performed fter visiting nd collecting smples from three trditionl utter mnufcturers nd one lrge scle utter producer. During these visits it ws possile to oserve the utter mking processes nd find out wht the trditionl utter production relly ment.. Mterils nd methods. Butter smples Trditionl utter smples were collected from three mnufcturers (A, B nd C). Five trditionl utters were collected. Three different utters from single industril utter mnufcturer (D) were lso collected. Trditionl utters were mde of different types of crem. Unslted nd slted utters were collected nd slted utters contined different levels of dded slt. One industril utter contined dditives tht ccording to the mnufcturer were imed to give the product sensory trits of the utter mde of cultured crem. Butter chrcteristics were summrised in Tle. Tle - Chrcteristic of collected utter smples Smple Fctory Declred slt Crem type used for Other dditives identifiction content production AU A None Whey None BU B None Sweet* None BS B.5% Clotted None CS C.5% Sweet* None CS C % Whey nd Sweet* None DU D None Sweet* None DU D None Sweet* Lctic Acid nd Dicetyl DS D % Sweet* None *Sweet refers to psteurised, uncultured crem. Sensory nlysis Butter smples were stored for h t C (ISO 935-, 9), nd were sujected to the sensory evlution right fter removing from this refrigerted storge. Sensory qulity of the utters ws evluted using pnel of 5 untrined volunteers, instructed on how to perform the utter nlysis directly prior to the test. The test ws designed ccording to ISO 935-:9. Volunteers were requested to ssess following trits of ech utter: odour, ppernce, flvour nd consistency in given order, using 5 point scle. The scle ws composed of following scores - For dislike very much - For dislike little 3- For not sure - For like little 5- For like very much Bsed on the results of the sensory nlysis n overll score for ech utter ws clculted. This score ws otined y multiplying score for ech ssessed trit y n ssigned importnce coefficient nd summing the result of ll the trits up in single score. The ssigned importnce coefficients for individul trits were sed on previously pulished work on sensory evlution of the utter (Czechowsk-Liszk, 5). The coefficients were individully for: odour-.3, ppernce-., flvour-. nd consistency-..
The volunteers were lso sked to rte utter smples from the one they liked est to the one they liked the lest. It ws possile to ssign sme score to multiple utter smples if they were of similr desirility. Slted nd unslted utters were evluted seprtely, strting with unslted utters..3 Chemicl nlyses Peroxide nd cid vlues of utter fts were evluted following procedures descried y Kruse (8) with some modifictions. Butter oil ws seprted from utter y liquidising smples in 5 C wter th for less thn min nd susequently centrifuging t 5 rpm for 5 min in MiniSpin Plus centrifuge with F-5-- rotor in.5. ml tues (Eppendorf, Hmurg, DE). Oil from -3 tues ws comined in single conicl flsk of known weight for cid or peroxide vlue evlution. Then smples were weighted nd nlysed s descried elow. In preprtion for titrimetric peroxide vlue evlution utter oil smple ws dissolved in chloroform ( ml). Then glcil cetic cid (5 ml) nd freshly prepred sturted potssium iodide solution (ml) were dded nd the smple ws shken y hnd for min, then kept in the drkness for exctly min. Susequently deionised wter (75 ml) nd freshly prepred % strch solution (out ml) were dded. Thus prepred smple ws titrted with. M sodium thiosulfte till lue colour disppered. Peroxide vlue ws clculted ccording to the eqution. PV = V st M st m Where: PV- peroxide vlue (meq of peroxide/kg of oil) V st- volume of sodium thiosulfte used for titrtion (ml) M st- molrity of sodium thiosulfte m- weight of the smple (g) For the cid vlue determintion solid oil smples were riefly (< min) heted in 5 C wter th till liquid, then dissolved in 5 ml of : (v:v) mixture of solute ethnol nd diethyl ether with dded ml of % phenolphthlein indictor nd neutrl ph (djusted with. M sodium hydroxide). Smple ws then titrted with.5 or.5 M NOH till development of pink colour tht ws stle for minimum of min. Acid vlue ws clculted ccording to the eqution. AV = V sh M sh m Where: AV- cid vlue (NOH/g of oil) V sh- volume of sodium hydroxide used for titrtion (ml) M sh- molrity of sodium hydroxide
m- weight of the smple (g) Peroxide nd cid vlues were estimted for fresh nd stored for 8,, 8 nd weeks smples t 5 C. Term fresh utter refers to up to week old utter supplied y the fctories nd stored in frozen conditions to the dy of nlysis. Moisture nd ft content were evluted using grvimetric method. To mesure moisture contentweight loss fter drying, smples were dried for h t 3 C. Dry smples were then sujected to ft extrction using Soxtec System HT 3 Extrction Unit (Tector LTD, Denton, Mnchester, UK) with : v:v mixture of dietyl nd petroleum ethers. Ft content ws clculted sed on the extrct weight. Wter ctivity ws mesured using AquL Series 3 TE wter ctivity meter (Lcell LTD, Four Mrks, Alton, UK). Averge temperture of the smples during this mesurement ws C. All nlyses were performed in triplicte.. Butter colour ssessment Colour mesurements performed y mens of Chrommeter CR- (Konic Minolt, Tokyo, Jpn) colorimeter. Vlues for lightness (L*) green to red (*) nd lue to yellow (*) s specified y Interntionl Commission on Illumintion (Interntionl Commission on Illumintion, ) were otined. For the purpose of this study only L* vlues were evluted..5 Microiologicl nlyses Butter smples were prepred for microiologicl nlyses following ISO 6887-3 (ISO 6887-, 3) method. Prior to enumertion, utter smples were suspended in Buffered Peptone Wter ( g of smple in 9 ml of liquid) nd homogenised t 5 rpm for min in Stomcher Circultor (Sewrd LTD, Worthing, UK). Otined homogente (net or diluted s necessry) ws then sujected to microiologicl nlysis. All smples were nlysed for initil contmintion (within week from production) with totl vile count (TVC), Enteroctericee, Pseudomons spp. nd yest nd moulds. All of these nlyses with exception of Enteroctericee were lso crried out for smples stored t 5 C for 8,, 8 nd weeks. Enteroctericee nd TVC were enumerted ccording to ISO 58- nd ISO 833-3 methods respectively. Pseudomons spp. were enumerted on Cetrimide-Fucidin-Cephlosporin Pseudomons gr nd yest nd moulds were enumerted on Rose Bengl Chlormphenicol gr. Both orgnisms were enumerted fter medi incution t 5 C for 8 (Pseudomons spp.) nd hours (yest nd moulds). All microiologicl medi were purchsed from Oxoid (Altrinchm, UK). All microiologicl nlyses were conducted in triplicte..6 Wter dispersion Butter smples prior to wter dispersion evlution were kept for h t 3 C (s recommended y stndrd (IDF A, 989)) nd then cut with thin knife. An indictor pper (Dysperwody, Llct, Olsztyn, PL) ws then ttched to freshly uncovered utter surfce, gently pressed nd removed to red results. Indictor ppers used in the study were infused with colournt tht stined the pper when in contct with moisture of the utter. The size nd totl re of the droplets visile on the pper were then used to ssess the wter dispersion in the utter s specified y the mnufcturer of the indictor ppers nd given in Tle.
Tle - Wter dispersion in utter ccording to Llct Result shown on indictor pper Score Approximte size of the moisture drops (mm) Approximte re of the pper covered y the moisture drops Very d 3 to 8 % Bd to 3 % Sufficient.3 to 5% Good No moisture drops visile.7 Sttisticl nlysis Sttisticl nlyses were performed using IBM SPSS Sttistics version. Significnce level for ll of the tests ws p<.5. Different sttisticl tests were used for different dt sets. Tests to which dt were sujected were mentioned in cptions to figures or tles where the results were summrised nd in text where pproprite. Men increse or growth rtes were given s slope of liner fit to the microiologicl growth curves (in log cfu g - ) or evolution of peroxide nd cid vlues mesured over weeks of storge. Microiologicl counts under the limit of detection were included into the sttisticl nlyses nd presented results s hlf of the vlue of the nlyticl detection limit expressed s cfu g -. The log cfu g - vlues of the detection limit were then.7 for Pseudomons spp., TVC nd yest nd moulds nd.7 for Enteroctericee. 3. Results nd discussion 3. Production process of trditionl nd industril utters, defining trditionl production Generlly production of utters in trditionl nd industril wy vried in production scle nd the wy utters were churned nd formed. The trditionl churning required filling nd emptying the churner fter ech utter production tch wheres the industril churning ws process in which crem ws continuously nd utomticlly spirted to the churner. In n industril production the churned utter did not require wshing with wter, conversely to the trditionl production. This nd full utomtion of industrilly produced utter were mking the whole process much shorter ( to 6 times) compred to the trditionl utter mking. Butter forming in the trditionl set-up ws done y hnd (AU, BU, BS), semi (CS) or full-utomtion (CS) while in the industril process it ws fully utomted. Two out of three trditionl utter mnufcturers were using wooden scotch hnds for utter forming (A nd B), ut churners in ll estlishments were mde of esily wshle mterils, such s luminium or stinless steel. Temperture of finished utter product prior to forming ws similr in etween industril nd trditionl processes (out 5 C), wheres crem storge temperture nd time depended on the individul mnufcturers. Further detils of the utter mnufcturing were summrised in
Tle 3.
Tle 3- Butter production prmeters Production step Crem storge temperture ( C) Crem storge time (dys) Mximum volume of the crem for single churning (l) Temperture of the crem dmitted to the churner ( C) Churning time (min) Buttermilk seprtion Mixing slt nd dditives Approximte durtion of the process (min) Butter temperture prior to forming ( C) Butter forming Mnufcturer A B C D 5 5 5 8 Up to 3 Up to Up to 5 if stored in silo nd up to if stored in pllecon Up to 35 Continuous production - 5 8-7 -5 5 Not specified Drined y opening churner s ottom vlve, utter is then wshed with wter In churner while kneding Drined y opening churner s ottom vlve, utter is then wshed with wter In churner while kneding Drined y opening churner s ottom vlve, utter is then wshed with chilled wter Upon the trnsfer to seprte mixing unit equipped with stirrer 9 to 5 5 9-5 5-6 5 g portions re weighted nd rolled y hnd, then formed using 5 g portions re weighted nd rolled y hnd, then formed using -5 g portions re utomticlly extruded nd cut y hnd or utomticlly Drined y ppliction of filters on wlls of the churner nd susequent two comprtments of the system where utter is mixed nd kneded, dditionlly first of these two comprtments pplies vcuum to help drining uttermilk Within one of the system s comprtments 5 kg portions utomticlly extruded
Production step Pckging Climed shelflife (weeks) Mnufcturer A B C D wooden scotch hnds wooden scotch hnds By hnd in By hnd in cellophne cellophne By hnd or utomted in comined lminte mteril Automted in crdord ox lined with polyethylene g 8 for unslted nd for slted utter 3. Sensory qulity of trditionl nd industril utters Results of sensory ssessment of trditionl nd industrilly produced utters were shown in Figure. Sttisticlly significnt differences for flvour nd odour s well s overll score were found in etween unslted utters. Trditionl utter BU got highest scores while nother trditionl utter (AU) nd industrilly mde utter (DU) lowest. Another industrilly mde utter DU got slightly, ut not significntly lower score compred to BU. The most fvourly perceived unslted utter ws BU (8 out of 5 prticipnts nominted it s the one they liked the most) followed y AU ( nomintions), DU (3 nomintions) nd DU ( nomintions). Results of sensory ssessment of unslted utters suggested tht the trditionl mnufcture of sweet crem utter llowed to otin product tht ws perceived more fvourly y the consumers. Nevertheless ddition of lctic cid nd dicetyl within the industril mnufcture of sweet crem utter resulted in product tht ws perceived nerly s good. Trditionl whey crem utter (AU) received slightly lower scores compred to the trditionl sweet crem utter (BU) ut it ws perceived eqully good s two industrilly mde utters nd chosen second est liked y the consumers utter. Whey crem utters, such s AU re known to hve different odour compred to the sweet crem utters (Jinjrk et l., 6), hence difference in sensory ssessment of these two types of utter ws expected. The sensory nlysis results indicted tht the whey crem utter is lso fvourly perceived y the consumers nd y some even more fvourly thn sweet crem utter. No differences in odour, flvour nd overll score etween slted utters were noted. Generlly this result indicted tht the ddition of slt msks differences tht cn e perceived due to use of different crem or process for utter production. One smple (BS) scored significntly lower for the ppernce thn the other smples. This smple ws chrcterised with drker colour (lower L* vlue) thn the other smples (see Figure ) nd hd visile drops of ccumulting moisture. Intense colour of the utter my e perceived s positive trit s drker utters tend to hve higher content of nturl colours such s crotene nd lycopene. These colours re ssocited with feed of the cows. Fresh grss fed cows will produce milk with higher content of the nturl colours thn hy or dry feed fed cows (Nozière et l., 6). Hence, utter mde of such milk my e perceived s more desirle y the consumers. Trditionl utters were generlly chrcterised y drker colour compred to the industril utters (see Figure ), nevertheless their ppernce ws not scored ny differently with exception of BS smple. Hence, it is likely tht the moisture visile on the BS utter ws the reson for lower scores for ppernce. The most fvourly perceived slted utter ws industrilly produced DS (7 out of 5 prticipnts nominted it s the one they liked the most) followed y ll trditionlly mde utters CS (6
nomintions), CS (3 nomintions) nd BS ( nomintions). Two est perceived utters hd higher dded slt level (%) compred to two utters with lower desirility (BS hd.5% nd CS.5%). Hence, it cn e concluded tht the slt level in slted utters is responsile for sensory desirility... 5 3 AU BU DU DU. 5 3 BS CS CS DS Figure - Results of the sensory evlution of unslted () nd slted () utters in 5 point scle. Error rs represent stndrd errors, while columns men scores. Different letters ove different columns elonging to the sme utter trit mrk sttisticlly significnt difference
L* score 9 9 88 c df f f 86 8 8 cd e e 8 78 76. 7 AU BU BS CS CS DU DU DS Figure - L* vlues for utter smples. Columns correspond to the men L* vlues while error rs to the stndrd error. Different letters ove different columns mrk sttisticlly significnt difference (p<.5 ANOVA nd Tukey post-hoc)..
3.3 Microiologicl qulity nd shelf-life of trditionl nd industril utters 3.3. Initil microiologicl contmintion Initil microiologicl contmintion on the smples of collected utter ws presented in Figure 3.. 7. 6 7 6 Log cfu g - 5 3 c d d d e 5 3 d c c c c c AU BU BS CS CS DU DU DS AU BU BS CS CS DU DU DS c. d. 3 5 Log cfu g - 3 AU BU BS CS CS DU DU DS AU BU BS CS CS DU DU DS Figure 3- Initil microiologicl lod on utter smples with respect to () TVC, () Pseudomons spp., (c) yest nd moulds, (d) Enteroctericee. Columns correspond to the men microiologicl count while error rs to the stndrd error. Different letters ove different columns in the sme grph mrk sttisticlly significnt difference (p<.5 ANOVA nd Tukey post-hoc). There were sttisticlly significnt differences in the initil microiologicl lod on the industril nd trditionl utters (MANOVA p<.5). The levels of microiologicl contmintion differed etween utter mnufcturers nd for TVC nd Pseudomons spp. were highest on smples from the mnufcturer B, followed y A, C nd D (MANOVA with ANOVA nd Tukey post-hoc p<.5). The highest count of ll the microorgnisms ws found on the smple BU, while smples CS, DU nd DU were chrcterised with lowest count, which ws elow the limit of detection. In remining smples Enteroctericee nd yest nd moulds were not detected, ut vrious levels of TVC nd Pseudomons spp. (rnging from under the limit of detection to 3.89 log cfu g - ) were found. Two out of 5 trditionl utters nd ll industril utters stisfied guided microiologicl criteri of no more thn 3 log cfu g - for TVC in freshly mnufctured utter (Adms nd Moss, 8). Only one smple (BU) hd TVC count ove 5 ut less thn 7 log cfu g - which ccording to Helth Protection Agency ws level tht my indicte need of further investigtion into sources of utter
contmintion (Helth Protection Agency, 9). This smple lso did not meet criterion of mximum dvisle level during utter s shelf-life for Enteroctericee of log cfu g - (Stnnrd, 997). Pseudomons spp. counts on this utter were lso high (5.8 log cfu g - ). This group of cteri multiplies rpidly compred to others in refrigerted diry (Fernndes, 9). Hence it cn e concluded tht either time or temperture of crem storge prior to utter production could hve compromised microiologicl qulity of this utter. 3.3. Microiologicl growth during refrigerted storge Full set of growth curves for ech of the tested microorgnisms on ech utter smple ws given in the supplementry informtion. Men growth rtes of microorgnisms on individul utters were presented in Figure..... Log cfu g - week -.6..8..6..8. AU BU BS CS CS DU DU DS AU BU BS CS CS DU DU DS c..5. Log cfu g - week -.5..5 AU BU BS CS CS DU DU DS Figure - Men growth rte of microorgnisms on utter stored t 5 C for weeks. () TVC, () Pseudomons spp., (c) yest nd moulds. All three types of microorgnisms were growing fstest on BS utter. The lowest men growth rte of TVC ws noted for DU utter, Pseudomons spp. for DU nd DS utters nd yest nd moulds for DS utter. Industril utters hd on verge lower men growth rtes compred to trditionl utters: TVC (.3 nd. respectively), Pseudomons spp. (on verge. nd.9 respectively) nd yest nd moulds (on verge. nd.3 respectively). Nevertheless, due to smll numer of smples significnt difference etween growth rtes of microorgnisms on trditionl nd industril utters ws only detected for yest nd moulds (t-test, p<.5). Growth curves of the microorgnisms on trditionl nd industril utters were presented in Figure 5.
. 7. 6 Log cfu g - 5 3 5 5 5 7 6 5 3 5 5 5 c. 7 Log cfu g - 6 5 3 trditionl industril 5 5 5 Figure 5- Growth curves of () TVC, () Pseudomons spp. nd (c) yest nd moulds on trditionl nd industril utters stored t 5 C.. Points correspond to the men microiologicl count while error rs to the stndrd error. Different letters ove different lines in the sme grph mrk sttisticlly significnt difference (p<.5 repeted mesures ANOVA). During the whole storge period industril utters hd significntly lower numer of ll microorgnisms compred to the trditionl utters. There were lso significnt differences in microorgnisms levels etween utters from different mnufcturers (see Figure 6).
.. Log cfu g - 8 6 c d 8 6 5 5 5 5 5 5 c. 8 Log cfu g - 6 c d A C B D 5 5 5 Figure 6- Growth curves of () TVC, () Pseudomons spp. nd (c) yest nd moulds on utters from different mnufcturers. Points correspond to the men microiologicl count while error rs to the stndrd error. Different letters ove different lines in the sme grph mrk sttisticlly significnt difference (p<.5 repeted mesures ANOVA nd LSD post-hoc). The highest microorgnism counts over the storge period were found on smples from the mnufcturer B, wheres lowest on the smples from the mnufcturer D. Smples from the mnufcturer C hd significntly higher TVC levels compred to smples from the mnufcturer D ut very similr Pseudomons spp. nd yest nd moulds levels for mjor prt of the storge. Addition of slt to utter is fctor tht should slow down or even eliminte growth of certin microorgnisms (Ledench nd Mrshll, 9). Here we hve oserved tht growth curves of TVC nd yest nd moulds on slted nd unslted utter were very similr (see Figure 7 nd c). The only group of microorgnisms tht presented slightly higher levels on unslted compred to slted utters ws Pseudomons spp., however given high vriility of the results this difference ws not significnt (Figure 7 c). A potentil explntion of this result my e non-uniform slt distriution in the utter (Fernndes, 9).
. Log cfu g - 7 6 5 3. 6 5 3 5 5 5 5 5 5 c. Log cfu g - 6 5 3 unslted slted 5 5 5 Figure 7- Growth curves of () TVC, () Pseudomons spp. nd (c) yest nd moulds on unslted nd slted utters stored t 5 C. Points correspond to the men microiologicl count while error rs to the stndrd error. Different letters ove different lines in the sme grph mrk sttisticlly significnt difference (p<.5 repeted mesures ANOVA). Summrising, otined results suggested tht growth of the microorgnisms on utter depended on the individul mnufcturing process ut not on whether the utter ws slted or unslted. Trditionl wy of utter mnufcture seemed to result in utters tht were more prone to microiologicl spoilge compred to industrilly mnufctured utters. However, n exmple of mnufcturer C shown tht t lest in regrds to yest nd moulds nd Pseudomons spp. this difference my e minimised. 3.3.3 Ft rncidity Full set of curves showing evolution of peroxide nd cid vlues over week storge t 5 C for oil seprted from ech utter smple ws given in the supplementry informtion. Peroxide nd cid vlues for fresh utter oils were summrised in Figure 8.
..8. mg of NOH g -.7.6.5..3.. meq kg -..8.6.. c d ce e d d ce AU BU BS CS CS DU DU DS AU BU BS CS CS DU DU DS Figure 8- Initil cid () nd peroxide () vlues of utter oil. Columns correspond to the men vlue while error rs to the stndrd error. Different letters ove different columns in the sme grph mrk sttisticlly significnt difference (p<.5 ANOVA nd Tukey post-hoc). Acid vlues of utter oils rnged from.8 to.6 mg of NOH g - with lowest for DS nd highest for BU smple. With exception of BU ll other utter oils were chrcterised with not sttisticlly significntly different cid vlues elow.57 mg of NOH g - which ws n dvised mximum for utter oil (FAO nd WHO, 7) (converted from cited puliction to units used in this work using clcultion given in (O Keefe nd Pike, )). Peroxide vlues rnged from. to.9 meq kg - with highest for DU nd lowest for CS smple. Although some sources gve n indiction tht peroxide vlues ove.3 meq kg - in utter oil should e ccompnied with detectle off-flvours (Munro et l., 998), no such off flvour ws detected in ny of the fresh utters during sensory nlysis. There were sttisticlly significnt differences in peroxide vlues of different utters however these did not seem to depend on whether the utter cme from trditionl or industril production. Mnufcturer C provided utter smples which oil ws chrcterised with significntly lower peroxide vlue compred to oil from utters produced y three other mnufcturers (on verge. meq kg -, ANOVA nd Tukey post-hoc, p<.5). The three remining mnufcturers including industril utter producer (D) provided utter smples with not significntly different to ech other peroxide vlues of oil (rnging on verge from.5 to.8 meq kg - )...8..7 mg of NOH g - week -.6.5..3.. AU BU BS CS CS DU DU DS meq kg - week -.35.3.5..5..5 AU BU BS CS CS DU DU DS Figure 9- Men increse rtes of cid () nd peroxide () vlue in oil seprted from utter stored t 5 C for weeks.
The cid vlue incresed fstest in BS nd slowest in DU utter oil, wheres peroxide vlue fstest in CS nd slowest in BU utter oil. High nd low increse rtes for peroxide vlues in oth industril nd trditionl utter oils were found. Hence it could e concluded tht trditionl utter mking did not impct increse rtes of peroxide nd cid vlues. Supporting this conclusion, no significnt differences in etween evolution of peroxide vlue over weeks of storge t 5 C ws found in etween oils from trditionl nd industril utters (see Figure ). Slightly, ut significntly higher cid vlues for oils in stored trditionl compred to industril utters were found. Nevertheless, comprison of cid vlue evolution in the smples otined from different mnufcturers reveled tht only mnufcturer B supplied utter smples with more rpidly incresing cid vlue compred to three other mnufcturers (see Figure ). In smples from the mnufcturer A peroxide vlue incresed more rpidly compred to smples from mnufcturers C nd D. However, evolution of peroxide vlue in oil from utters produced y mnufcturers B nd C did not significntly differ from one in oil from utters provided y mnufcturer D....8.6.. mg of NOH g - meq kg -..8.6.. 5 5 5 5 5 5 trditionl industril Figure - Evolution of cid () nd peroxide () vlue in oil of trditionl nd industril utters stored t 5 C. Points correspond to the men microiologicl count while error rs to the stndrd error. Different letters ove different lines in the sme grph mrk sttisticlly significnt difference (p<.5 repeted mesures ANOVA).
..8.. mg of NOH g -.6...8.6.. All curves elow meq kg -..8.6.. A B C D 5 5 5 5 5 5 Figure - Evolution of () cid, () peroxide vlue in oil of utters from different mnufcturers. Points correspond to the men microiologicl count while error rs to the stndrd error. Different letters ove different lines in the sme grph mrk sttisticlly significnt difference (p<.5 repeted mesures ANOVA nd LSD post-hoc).....8 mg of NOH g -.6.. meq kg -.8.6.. 5 5 5 5 5 5 unslted slted Figure - Evolution of () cid nd () peroxide vlue in oils of unslted nd slted utters stored t 5 C. Points correspond to the men microiologicl count while error rs to the stndrd error. Different letters ove different lines in the sme grph mrk sttisticlly significnt difference (p<.5 repeted mesures ANOVA). Evolution of cid nd peroxide vlues ws not significntly different etween slted nd unslted utters (see Figure ). Summrising, otined results suggested tht the ft rncidity progressed t very similr rte in trditionl nd industril mnufctured utters. Additionlly initil qulity of the utter oil ws similr in etween two utter types. Addition of slt did not hve n impct on the rncidity progress. 3.3. Sensory ssessment of shelf-life Appernce nd odour of utters during storge t 5 C for weeks were evluted. Imges of the smples tht were frozen t the different stges of the shelf-life were included in Figure 3. Best qulity during tested period of time (no off odours nd no visile growth of microorgnisms for whole
shelf life tril) ws noted for utters DU, DS, CS nd CS. Butter DU y week developed very slight visile microorgnism growth ut no off odour. Butter AU lso hd slight visile growth y the week nd dditionlly slight cheesy off odour. Butters BU nd BS were visully nd odour-wise not cceptle efore week 8 of storge, they hd visile growth of colonies nd distinct cheesy offodour. Where found, visile microorgnism growth ws only present on the surfce nd not inside of the utter locks (see Figure 3). All the mnufcturers judged the shelf life of offered utters correctly s neither of the smples spoiled efore termintion of the stted shelf life. Spoilge ws ccompnied y TVC counts of.98 to 8.8 log cfu g -, Pseudomons spp. counts of under the limit of detection up to 8.8 log cfu g - nd yest nd moulds counts of.89 to 7. log cfu g - s well s peroxide vlues of. to. meq kg - nd cid vlues of. to.9 mg of NOH g -. Lrge rnge of these vlues indicted tht fctors other thn ft rncidity development nd growth of the microorgnisms were lso contriuting to the utter spoilge. The fct tht the growth of microorgnisms strted to e visile t much different concentrtions in different smples could e lso consequence of smpling nd visile ility of microorgnisms to grow only on the surfce of the utter. Another importnt fctor for longevity of the utter is wter dispersion (Fernndes, 9). Results of wter dispersion test were descried in section 3..
Figure 3- Appernce of the utters during shelf life tril, smples on the left side of ech picture represent utter surfce nd on the right- utter cross section frgments. Numers correspond to the week of the storge. Results for smple DS were not given s they were very similr to smple DU.
3. Wter dispersion in utter The results of wter dispersion nlysis were presented in Figure. All trditionl utters with exception of CS contined lrge droplets of wter (lrger thn 3 mm to ove cm) tht covered lrge surfce prt of the indictor pper (more thn %). Butter BS ws chrcterised with the worst wter dispersion in this group nd AU est. Industril utters nd CS, ccording to the scle given y the mnufcturer of the indictor ppers used in this study, hd n cceptle wter dispersion with smll droplets of wter (-3 mm) covering up to 5% of the pper s surfce. Figure - Wter dispersion in utter
3.5 Chemicl ssessment of trditonl nd industril utters Tle - Chemicl composition nd wter ctivity of studied utters men vlues ± stndrd errors. Different letters in different rows of sme column mrk sttisticlly significnt difference (p<.5 ANOVA nd Tukey post-hoc). Butter smple Moisture % Ft % Wter ctivity AU.6 ±. 86. ±..96 ±. BU. ±. 86.5 ±..98 ±. BS 3.6 ±. cd 85. ±..8 ±. c CS.6 ±.3 d 86.8 ±..73 ±. CS 3.6 ±. cd 85.6 ±.5.73 ±. DU.3 ±.3 ce 8.7 ±..97 ±. DU 5.9 ±.5 e 83.9 ±.3 c.98 ±. DS. ±. de 83. ±.3 c.7 ±. There were significnt differences in moisture nd ft content etween the utters (see Tle ). Highest moisture content ws found in DU while lowest in BU utter. Butter CS ws chrcterised with highest mount of ft, while utter DS with lowest. Products from trditionl utter mnufcturers ll contined similr mount of ft nd moisture (ANOVA nd Tukey post-hoc p>.5). Further dt nlysis indicted tht industrilly mde utters contined significntly (t-test, p<.5) more moisture (on verge 5. %) nd less ft (on verge 8.3 %) compred to trditionl utters (on verge 3. nd 86.5 % respectively). This outcome is not surprising since, within the industril process, mnufcturer ws le to monitor nd control the mount of wter in the utter, while no similr control ws oserved in the trditionl production. The wter ctivity differed in etween utters (see Tle ). This difference ws depending on slt content in ech utter. Butters with % slt content hd wter ctivity rnging from.7 to.73 while unslted utters from.96 to.98. Butter with.5 % slt content hd wter ctivity of.8. All slted utters hd wter ctivity on level preventing growth of cteri, ut llowing growth of yests nd moulds (Fontn, 7).. Conclusion The trditionl nd industril utter mnufcture differed considerly. The trditionl utter ws mde using tch churning process while industril utter using continuous process. Both these processes were well descried y Wiley (). Additionlly trditionl utter mnufcturing involved in most of cses (four out of five utters) hnd forming nd pckging of the utter. Due to the production scle the industril utter ws mde sed on milk from mny different suppliers while trditionl mnufcturers used predominntly milk from their own cows (in this study two out of three mnufcturers did so). This study found tht the trditionl utters my e indeed preferred y the consumers to industrilly produced, high qulity utters. This hypothesis ws confirmed sed on unslted utters. Nevertheless, sensory qulity of the industril utter could e considerly improved y ddition of lctic cid nd dicetyl, mtching high sensory scores otined y the trditionl utter.
Slted utters were ssessed y the sensory pnel eqully well regrdless of whether they were industrilly or trditionlly produced. The sensory cceptnce of slted utters seemed to improve with increse of the slt level. Sensory evlution of the trditionl utters lso indicted tht ppernce of these products my e compromised y poor wter dispersion. The wter dispersion ws generlly worse for trditionl compred to industril utters (in four out of five cses). This fctor could influence growth of the microorgnism. Microorgnism growth indeed tended to e fster on trditionl compred to the industril utters. Fster growth of microorgnisms nd higher initil microiologicl lod found on some of the trditionl utter smples could result in shorter shelf life of these products. Nevertheless this ws not the cse for ll the smples. Only two out of five tested trditionl utters hd shorter shelf life (evluted sed on sensory cceptnce) compred to the industrilly produced utters. Both of these utters were produced y the sme mnufcturer. One of these utters crried higher microiologicl counts compred to ll the other nlysed utters. Another one ws chrcterised y very poor wter dispersion nd more rpid increse of microorgnisms popultion over the storge period compred to ll other nlysed utters. It is hence proposed tht the mnufcturer could prolong shelf life of the utter y the improvement of the wter dispersion nd reduction of the initil microiologicl contmintion. Poor wter dispersion my e consequence of temperture during churning eing too high. At churning temperture higher thn 3 C high mount of utter ft is liquid nd kneding does not llow for working the moisture into the utter very well (Nielsen, 97). High microiologicl counts in sweet crem utter right fter production my e consequence of:. high microiologicl counts in crem,. utter nd crem hndling during nd post production nd 3. insufficient removl of the uttermilk. For n dequte removl of the uttermilk utter grins need to e wshed -3 times with drinking wter of good microiologicl qulity e.g. psteurised nd in volume corresponding to the volume of crem used for utter production (Berthold-Plut et l., 3; Nutrivitlity). Wshing improves microiologicl qulity nd shelf life stility of the utter y removl of residul uttermilk together with the microes nd reduction of nutritious proteins nd lctose, which would llow these microes to grow (Berthold-Plut et l., 3; Spreer, 998). Wshing my lso e used to control temperture of utter grins prior to kneding (importnt for pproprite wter dispersion) nd is known to limit lipolysis in the utter (Berthold-Plut et l., 3). Nevertheless, some uthors rgue tht wshing reduces flvour creting compounds, resulting in lower sensory desirility (Spreer, 998). Qulity of ft in fresh nd stored trditionl utters ws very similr to industril utters evluted in this study. The only exception were utters from trditionl mnufcturer B. In these utters cid vlues incresed more rpidly thn in others. Summrising, this study confirmed consumer perception tht trditionl utters my e chrcterised y higher sensory desirility compred to the industril utters. It did not miguously confirm clim tht the trditionl utters hve shorter shelf life thn industril utters. This presumption held true only for utters from one out of three trditionl mnufcturers. Further insight into individul processes of this mnufcturer could help improve shelf life of the offered utters. Acknowledgment This work received funding from The Frmer s Clu Chritle Trust, London nd University of Lincoln s, Ntionl Centre for Food Mnufcturing, Holech. Author would like to express grtitude to ll utter producers who greed to demonstrte their utter processing nd supply smples for this study.
5. References Adms, M. R. & Moss, M. O. (8) The Microiology of Food Preservtion, in Food Microiology 3rd Edition. RSC Pulishing. pp. 63 8. Autio, M. et l. (3) Consuming nostlgi? The pprecition of uthenticity in locl food production. Interntionl Journl of Consumer Studies. 37 (5), 56 568. Berthold-Plut, A. et l. (3) Przetworstwo mlek n poziomie gospodrstw. Aville from: http://cpsz.cdr.gov.pl/imges/site/pliki/wyd/3/pzetworstwo_mleko.pdf. Budkhr, Y. A. et l. () Microiology of Crem nd Butter. Encyclopedi of Food Microiology, Volume. 78 737. Chmpgne, C. P. et l. (99) Psychrotrophs in diry products: Their effects nd their control. Criticl Reviews in Food Science nd Nutrition. 3 (), 3. Czechowsk-Liszk, M. (5) Bdnie i ocen jkości różnych rodzjów msł dostępnego n rynku. Zeszyty Nukowe Akdemii Ekonomicznej w Krkowie. 67877 87. FAO & WHO (7) Codex Alimentrius, Milk nd Milk Products. Aville from: ftp://ftp.fo.org/docrep/fo//387e/387e.pdf (Accessed 6 April 6). Fernndes, R. (9) Butter nd Diry Spreds, in Fernndes R. (eds.) Microiology Hdook: Diry Products. Letherhed Pulishing. pp. 9 6. Fernndes, R. (9) Liquid Milk Products, in Fernndes R. (eds.) Microiology Hdook: Diry Products. Letherhed Pulishing. pp.. Fontn, A. J. (7) Appendix D: Minimum Wter Activity Limits for Growth of Microorgnisms, in Gustvo V. Bros-Cánovs et l. (eds.) Wter Activity in Foods. Blckwell Pulishing Ltd. pp. 5 Helth Protection Agency (9) Guidelines for Assessing the Microiologicl Sfety of Redy-to-Et Foods Plced on the Mrket. Aville from: https://www.gov.uk/government/uplods/system/uplods/ttchment_dt/file/3636/g uidelines_for_ssessing_the_microiologicl_sfety_of_redy-toet_foods_on_the_mrket.pdf (Accessed 6 April 6). Higgins, R. & Ptterson, A. (3) Our Butter hs Become the Tost of the Top Tles. The Sundy Times. Aville from: http://www.ernethyuttercompny.com/wordpress/wpcontent/uplods/3//ldaernethy.pdf (Accessed 6 April 6). IDF A (989) Butter - Determintion of wter dispersion vlue. Interntionl Commission on Illumintion () Technicl Report, Colorimetry 3rd Edition. Aville from: http://cie.mogi.me.hu/cie_rch/kee/div/tc8.pdf (Accessed 7 April 6). ISO 833- (3) Microiology of the food chin -- Horizontl method for the enumertion of microorgnisms -- Prt : Colony count t 3 degrees C y the surfce plting technique.
ISO 6887- (3) Microiology of food nd niml feeding stuffs -- Preprtion of test smples, initil suspension nd deciml dilutions for microiologicl exmintion -- Prt : Specific rules for the preprtion of met nd met products. ISO 58- () Microiology of food nd niml feeding stuffs -- Horizontl methods for the detection nd enumertion of Enteroctericee -- Prt : Colony-count method. ISO 935- (9) Milk nd milk products- Sensory nlysis- Prt : Recommended methods for sensory evlution. Jinjrk, S. et l. (6) Sensory, Functionl, nd Anlyticl Comprisons of Whey Butter with Other Butters. Journl of Diry Science. 89 (7), 8. Kruse, A. J. et l. (8) The effect of refrigerted nd frozen storge on utter flvor nd texture. Journl of Diry Science. 9 (), 55 65. Ledench, L. H. & Mrshll, R. T. (9) Microiologicl Spoilge of Diry Products, in Sperer W. H. nd Doyle M. P. (eds.) Food Microiology nd Food Sfety, Compendium of the Microiologicl Spoilge of Foods nd Beverges. Springer. pp. 68. McEchern, M. G. et l. () Thinking loclly, cting loclly? Conscious consumers nd frmers mrkets. Journl of Mrketing Mngement. 6 (5 6), 395. Munro, D. S. et l. (998) Concentrted Milkft Products, in Erly R. (eds.) Technology of Diry Products, nd Edition. Blckie Acdemic & Professionl. pp. 98. Neves, P. & Lngridge, E. W. (998) Lortory control in milk product mnufcture, in Erly R. (eds.) The Technology of Diry Products, nd Edition. Blckie Acdemic & Professionl. pp. 368-. Nielsen, V. H. (97) Curing the texture defects in utter. Americn Diry Review. 338 9, 9. Nozière, P. et l. (6) Vritions in Crotenoids, Ft-Solule Micronutrients, nd Color in Cows Plsm nd Milk Following Chnges in Forge nd Feeding Level. Journl of Diry Science. 89 (7), 63 68. Nutrivitlity (n.d.) Śmietnk, śmietn, msło. Aville from: http://www.nutrivitlity.pl/index.php?option=com_content&view=rticle&id=&itemid= (Accessed 6 April 6). O Keefe, S. F. & Pike, O. A. () Chemicl Properties nd Chrcteristics of Foods- Ft Chrcteriztion, in Nielsen S. S. (eds.) Food Anlysis. Springer. pp. 39 6. Spiller, K. () It tstes etter ecuse consumer understndings of UK frmers mrket food. Appetite. 59 (), 7. Spreer, E. (998) Butter Mnufcture, in Spreer E. (eds.) Milk nd Diry Product Technology. Mrcel Dekker. pp. 3. Stnnrd, C. (997) Development nd use of microiologicl criteri for foods. Food Science nd Technology Tody. (3), 37 77. Wiley, A. () Microiology of Crem nd Butter, in Roinson R. K. (eds.) Diry Microiology Hndook, The Microiology of Milk nd Milk Products. Wiley. pp. 3 7.