J Sci Food Agric 1997, 75, 433È441 Formation of White Spots in the Shell of Raw Shrimps During Frozen Storage. Seasonal Variation and Effects of Some Production Factors Anni Mikkelsen,1* Birgitte RÔnn2 and Leif H Skibsted1 1 Department of Dairy and Food Science, Royal Veterinary and Agricultural University, Rolighedsvej 30, DK-1958 Frederiksberg C, Denmark 2 Department of Mathematics and Physics, Royal Veterinary and Agricultural University, Thorvaldsendsvej 40, DK-1871 Frederiksberg C, Denmark (Received 5 August 1996 ; revised version received 11 April 1997 ; accepted 18 April 1997) Abstract : During frozen storage of raw pink shrimps, Pandalus borealis, calcium carbonate tends to precipitate in the exoskeleton, giving the shrimps a spotted appearance. Seasonal variations in the formation of white spots and the inÑuence of shrimp size, production time, chemical treatment, and storage temperature on white spot formation was determined. A signiÐcant e†ect (P \ 0É003) of day of catch was observed and shrimps caught in the summer period showed a greater tendency of white spot formation than shrimps caught during the rest of the year. Treatment in a sulphite, a phosphate or a phthalate solution prior to freezing retards calcium carbonate crystallisation (P \ 0É05) while treatment in a borax solution promotes crystallisation (P \ 0É05). Prolonged time from catch to chemical treatment and increased storage time increase the risk of white spot formation. Storage temperature has a signiÐcant e†ect on white spot formation for which process the rate at di†erent temperatures was described by the Arrhenius equation with a high energy of activation of 90 kJ mol~1. The size (age) of the shrimps did not inÑuence white spot formation (P \ 0É69). From these Ðndings it is concluded that calcium carbonate precipitation resulting in white spots in shrimp shell can be retarded or possibly prevented by short production time, appropriate chemical treatment and very low storage temperature. J Sci Food Agric 75, 433È441 (1997) No. of Figures : 5. No. of Tables : 3. No. of References : 18 Key words : pink shrimp, Pandalus borealis, white spots, calcium carbonate crystallisation, frozen storage, shrimp size, chemical treatment, production time, storage temperature INTRODUCTION quality problem is formation of white spots on the shrimp shell a†ecting the appearance but not the taste or other eating qualities of the shrimp. Among the shrimp industry the problem is well known. However, to the best of our knowledge, no scientiÐc papers dealing with this topic has been published. Pandalus borealis belongs to the class Crustacea of the order Decapoda of the family Pandalidae (Shumway et al 1985) and, as for other decapods, the shell consists of chitin, protein and calcium salts (Stevenson 1985). In order to allow growth, the exoskeleton is periodically shed and replaced (Gnatzy and Romer 1984 ; Stevenson Raw shrimps are a very popular and rather expensive food product in Asia. Among other species, the pink shrimp, Pandalus borealis (KrÔyer 1838), is processed for this product. The shrimps are frozen shortly after catch and kept frozen until they reach the consumer. Raw shrimps with shell on are also frequently used for decoration of meals and the appearance of the shrimps is a very important quality parameter. A common * To whom correspondence should be addressed. 433 ( 1997 SCI. J Sci Food Agric 0022-5142/97/$17.50. Printed in Great Britain A Mikkelsen, B RÏnn, L H Skibsted 434 1985) and the physical properties of the shell therefore varies during the growth cycle. This might be of importance in relation to white spot formation. When the shrimps are caught and still alive no white spots are observed but after some time of frozen storage white spots can develop in the shell. During continued frozen storage the spots increase in size and consequently the quality of the product decreases. We have recently found that the white spots are crystals of calcite and vaterite, two forms of calcium carbonate, in a matrix of chitin (Mikkelsen et al 1997). Whereas the chemical composition of the white spots is now known, the conditions triggering the precipitation of calcium carbonate have not been identiÐed yet. From non-systematic observations, formation of white spots in the shell of frozen raw shrimps seems to follow a seasonal variation and to be a natural event that, to some extent, might be controlled during the production process and storage conditions. The usual production process of “frozen raw shell-on shrimpsÏ includes mechanical sorting according to size, treatment in a glazing-solution (chemical treatment), packaging in 1 kg plastic-coated cardboard boxes, freezing in a plate freezer and storage on board the trawler at c [25¡C until unloading. After unloading the shrimps are kept frozen during storage and transportation. The aim of the present work was to detect possible seasonal variations of white spot formation and to determine the inÑuence of the following production and storage parameters on white spot formation : shrimp size, production time, chemical treatment, and storage temperature. The inÑuence of shrimp size, reÑecting the age of the shrimp, might be considered a biological parameters as well. The work was conducted on board shrimp trawlers, and samples were taken from di†erent sites in the usual production and exposed to the desired treatments. Unfortunately, it was not always possible to conduct all planned sampling and treatments of shrimps on board trawlers. This is the reason for di†erent number of repeating of treatments. EXPERIMENTAL Shrimp samples were taken at di†erent sites at the usual production line on board Royal Greenland shrimp trawlers operating in the Davis Strait (between 60¡ and 72¡N and between 60¡W and the coast line of Greenland) and used for the investigations described below. After mechanical sorting according to size the shrimps were treated in a glazing solution for one minute. Immediately after this chemical treatment the shrimps were packed in 1 kg low-density polyethylenecoated cardboard boxes and frozen in a plate freezer to a core temperature of [30¡C or below controlled with an electronic thermometer with the sensor in the product. The shrimps were stored in a freezer storage at c [25¡C on board the trawler until unloading. In each experiment all samples were from the same catch. Seasonal variations Random sampling of raw shrimps from ten trawlers during 22 months was conducted to study the variation in white spot occurrence during the year and to study the dependence of white spot occurrence upon time from catch to inspection. These samples (in total 2096 1-kg boxes) were taken from the usual production after freezing and analysed shortly after unloading. As the trawlers were catching for up to 2 months, samples were stored for 1È70 days at [25¡C before being unloaded and analysed. E†ect of shrimp size Shrimps of the two most common size categories were used to investigate the inÑuence of size of the shrimps on formation of white spots. Eighteen 1-kg boxes of each shrimp size (small : 9È11 g, and large : 11È14 g) were taken out from the production line after freezing. The same number of samples were taken from three catches and examined after 26, 34, 39, 41, 45 and 48 days of storage, respectively. E†ect of chemical treatment Experimental chemical treatments were made in buckets with shrimps sampled at the production line prior to the normal chemical treatment : portions of 2È3 kg of shrimps were poured into 6 litres of the experimental “chemical solutionÏ. After 1 min the solution was poured through a strainer into another bucket, and the shrimps were drained for a few seconds, packed in 1-kg boxes and frozen. For each chemical treatment, 6 kg in total were prepared in the solution. Four individual series of chemical treatments were conducted, each repeated two or three times. All “chemical solutionsÏ (listed in Table 1) were made with sea water. Glaze-Nole, BL-7P (Shimakyu Chemical Co, Osaka, Japan) and Oxinon (Cosmo Techno Co, Kanagawa, Japan) are commercial products for chemical treatment of shellÐsh, and ingredients of these products are given in Table 2. Potassium phthalate, potassium dihydrogen phosphate, disodium diphosphate (Riedel-de Haen, Seelze, Germany), sodium tetraborate (Merck, Darmstadt, Germany), sodium sulphite, pentasodium triphosphate, and trisodium trimetaphosphate (Sigma, St Louis, MO, USA) were all of analytical grade. Solution pH was measured with a pocket pH meter (accuracy ^0É2). W hite spots in the shell of raw shrimps during frozen storage 435 TABLE 1 E†ect of chemical treatments on white spot formation in shell of raw frozen shrimps investigated in four experimental series ; “commercial productÏ, “pHÏ “sulphiteÏ, and “phosphatesÏ. Chemical treatmenta Solution pHb Indexc E†ect of commercial products Sea water Glaze-Nole (10 g litre~1) BL-7P (15 g litre~1) ] Glaze-Nole (10 g litre~1)d BL-7P (15 g litre~1) Oxinon (30 g litre~1)d Oxinon (30 g litre~1), 2 min Oxinon (30 g litre~1), 3 min Oxinon (30 g litre~1), 6 min Oxinon (60 g litre~1) 7É8 5É9 5É1 4É9 5É7 5É7 5É4 5É7 5É7 2É22a 1É72b 0É15c 0É11c 0É26c 0É30c 0É17c 0É12c 0É18c E†ect of pH Sea water BL-7P (15 g litre~1) Oxinon (40 g litre~1)d phthalate (0É02 M) phthalate (0É02 M) phosphate (0É02 M) phosphate (0É02 M) borax (0É005 M) borax (0É005 M) 7É7 5É4 5É5 4É5 4É7 5É4 6É1 7É0 7É6 2É98b 1É06c 0É99c 2É72b 2É52b 1É20c 1É38c 3É38a 3É36a E†ect of sulphite Sea water BL-7P (15 g litre~1) Oxinon (40 g litre~1)d sulÐte (0É016 M) ] phthalate (0É02 M) sulÐte (0É032 M)e ] phthalate (0É02 M) sulÐte (0É048 M) ] phthalate (0É02 M) 7É5 5É3 5É4 5É2 5É1 5É0 3É85a 1É84d 1É86d 3É11b 2É21c 1É51e E†ect of phosphates Sea water BL-7P (15 g litre~1) Oxinon (40 g litre~1)d Diphosphate (0É003 M)e Diphosphate (0É003 M)e ] phthalate (0É02 M) Tripolyphosphate (0É002 M)e Tripolyphosphate (0É002 M)e ] phthalate (0É02 M) Trimetaphosphate (0É001 M)e Trimetaphosphate (0É001 M)e ] phthalate (0É02 M) 7É3 5É4 5É4 5É3 5É4 7É1 5É5 7É5 5É4 3É79a 1É68c 1É77c 2É92ab 2É63b 2É95ab 3É26ab 3É40ab 3É64a a All solutions were made with sea water, and all treatments were for 1 minute when not otherwise indicated. b pH values are mean values of 2È3 experiments. c Estimated white spot index, (k ] a ) in equation 3 in the text. Within an experimental series, index i values with the same following letters are not signiÐcantly di†erent at the 5% level. d Common production solution. e Concentration as in 15 g litre~1 BL-7P solution. E†ect of production time The production time consist of two intervals, that is time from catch to chemical treatment and time from chemical treatment to freezing. Boxes of shrimps were randomly sampled at the production line immediately after chemical treatment and packaging in 1-kg boxes at the beginning of the production (Time-1 and Time-1a, see Table 3), and about 1 h later (Time-2). For each of the three treatments, three 1-kg boxes were used. A Mikkelsen, B RÏnn, L H Skibsted 436 TABLE 2 Ingredients of commercial products for treatment of shellÐsh according to product speciÐcations. All three products are produced and allowed for use in food in Japan. Product name Ingredients Concentration (g kg~1) Oxinon Sodium sulphite Sodium polyphosphate Sodium alginate “Natural substancesÏ 175 75 30 720 BL-7P L-Tryptophan Sodium metaphosphate Sodium D-tartrate Sodium polyphosphate Sodium citrate Sodium hydrogenpyrophosphate Sodium L-glutamate Sodium erythorbate Sodium sulphite 8 24 32 47 33 46 25 113 672 Glaze-Nole “Natural substancesÏ Time-1 and Time-2 samples were placed in the freezer immediately after packaging, while Time-1a samples were placed in the freezer at the same time as Time-2 samples. The experiment was repeated twice. E†ect of storage temperature Samples (312 1-kg boxes) were taken from the production line after freezing. After unloading, the samples were placed in freezers at [16¡C(^2¡C), [18¡C(^2¡C), [25¡C(^3¡C) and at [29¡C(^2¡C), respectively. The temperature of each freezer was measured by continuous temperature recordings throughout the investigation period. Samples were analysed several times during a storage period of 63 days, the samples stored at the lower temperatures observed during 1000 longest period (see Fig 4 below). To study the e†ect of a short storage period at high temperature, simulating temperature abuse during storage and transportation, on later development of white spots, samples were moved from [16¡C and [18¡C, respectively, to [25¡C at day 6, and stored at [25¡C for the rest of the storage period. The e†ect of temperature Ñuctuations were studied by moving samples back and forth between [29¡C and [25¡C 12 times with minimum 24 h between each movement. Visual examination All analyses of white spots in shrimp shell were made by visual inspection of the shrimps, and for each treatment three 1-kg boxes were analysed. Each 1-kg box was TABLE 3 E†ect of production time on white spot formation in shell of raw frozen shrimps treated in 0É40 g litre~1 Oxinon for 1 min. Production conditions Time-1 Time-1a Time-2 Indexa Chemical treatment at the beginning of the production and immediately placed in the plate freezer Chemical treatment at the beginning of the production and placed in the plate freezer 1 h later Chemical treatment after 1 h production and immediately placed in the plate freezer 1É65a 1É54a 2É15b a Estimated white spot index, (k ] a ) in eqn (3) in the text. Index values with the same following i letter (a or b) are not signiÐcantly di†erent at the 5% level W hite spots in the shell of raw shrimps during frozen storage thawed under tap water (15¡C ^ 1) for 10 min and inspected immediately. Depending on the size of white spots each shrimp were given a score ; 0 \ no white spots ; 1 \ very small white spots ; 2 \ small white spots ; 3 \ large white spots and 4 \ splotches of white spots, and for each 1-kg box a white spot index has been calculated index \ ; c N /N (1) j j where c is the score (0È4), N is the number of shrimps j j with score c , and N is the number of shrimps in the j box. Thus, “indexÏ, the empirical mean score of a 1-kg box, accounts for the degree of white spots, including both the number of shrimps with white spots as well as the size of the white spots. Statistics In the investigation of seasonal variation, the inspection was made shortly after unloading. White spots had developed in very few boxes and an assumption of normally distributed observations of the white spot index would not be reasonable. Therefore, the index was transformed into a binary response, valued 0 for index \ 0 and 1 for index [ 0. The transformed response was analysed by a logistic regression with the day of catch and the time between catch and inspection included as Ðxed covariates, and the day of inspection included as a random e†ect. The model is described by the formula ln[p /(1 [ p )] \ a ] b É cos(2n É day /365) i i i ] c É sin(2n É day /365) ] d É time ] C (2) i i time~i where p is the probability of box i to contain any i shrimps with white spots, and p /(1 [ p ) is odds of i i white spots for box i. Day is the day of catch and time i i is the time between catch and inspection for box i. The random variable C is the e†ect of the day of time~i inspection and is assumed to be independent and normally distributed. For more details see (McCullagh and Nelder 1989). Data from the three series of investigations concerning shrimp size, chemical treatment and production time were analysed by analysis of variance, with the relevant treatment, ie shrimp size, chemical treatment or production time, included as a Ðxed e†ect in the model. A considerable variation between experiments were observed, and accordingly random e†ects of the experiment as well as interaction between the experiment and the treatment were included in the model. index \ k ] a ] E ] (T E) ] e (3) ijk i j ij ijk where index is the observed index for box k given ijk treatment i in experiment j. k is 1È3 as three boxes were analysed for each treatment. The number of treatments, 437 i, varies between experimental series ; i is 1È3 for the experimental series of shrimp size, 1È9 for the experimental series of chemical treatment “commercial productsÏ, “phÏ and “phosphatesÏ, 1È6 in the experimental series of chemical treatment “sulphiteÏ (see Table 1), and 1È3 in the experimental series of production time (see Table 3). j is the number of repetitions of the experiment. k is the overall mean index, a is the e†ect of i treatment i, and E , (T E) are random e†ects of the j ij experiment and interaction between the experiment and the treatment, respectively. e is the measurement ijk error. In the shrimp size investigation, the shrimps were examined several times during the storage period, and the storage time (time between production and inspection) was included in the model as a Ðxed e†ect as well index \ k ] a ] c É time ] E ] (T E) ] e (4) ijk i ijk j ij ijk Before the analysis, the white spot index was transformed by arcsin()index) to achieve a homogeneous variance. Data from the experimental series concerning storage temperature were evaluated using a model where formation and growth of white spots during storage is described by a polynomium of third degree in the storage time with a rate depending on the temperature : arcsin()index ) \ a É (rate É time ) ] i temp i b É (rate É time )2 ] c É (rate É time )3 ] e (5) temp i temp i i where index is the observed index for box i, time is the i i storage time, and rate is the rate parameter corretemp sponding to the storage temperature of box i. a, b and c are constants, and e is the measurement error. The temi perature dependence of the rate was evaluated according to an Arrhenius type equation \ A É exp[[(E /R) É T ~1] (6) temp a where A is a constant, E is energy of activation, R is a the gas constant and T is the absolute temperature. All statistical calculations were made by the procedures MIXED and NLIN, and the macro %GLIMMIX in SAS version 6.10. rate RESULTS Variation in white spot occurrence during a period of 22 months is shown in Fig 1. Logistic regression analysis showed a signiÐcant e†ect (P \ 0É003) of the day of catch on the probability of white spot occurrence, and shrimps caught in the summer seasons had markedly increased incidence of white spots. The statistical analysis also showed a signiÐcant e†ect (P \ 0É001) of storage time (at c [25¡C) on white spot formation, as may be seen from Fig 2. 438 A Mikkelsen, B RÏnn, L H Skibsted Fig 3. Development in white spots during frozen storage of small (K) and large (=) shrimps. Each data point is a mean of 9 observations (3 experiments with each 3 observations). Fig 1. White spot occurrence in raw frozen shrimps as function of time of catch. Day of catch \ 0 is 1 January 1994. L, Observed mean odds of grouped data from random sampling during 22 months. …, Predicted odds of grouped data (same mean time of catch as for observed data), corrected for the groupÏs mean time between time of catch and inspection as for the observed odds. Full line, predicted odds, for time between catch and inspection being equal to mean time between catch and inspection (\24É3) for the entire data set. Fig 2. White spot occurrence in raw frozen shrimps as function of time (days) between catch and inspection. L, Observed odds of grouped data from random sampling during 22 months. …, Predicted odds of grouped data (same mean time between catch and inspection as for observed data). Full line, predicted odds, for time of catch being equal to mean time of catch (\335É2) for the entire dataset. Development in white spots during frozen storage of small and large shrimps is shown in Fig 3. The data conÐrm the e†ect of storage time, but no di†erence between the two sizes was observed (P \ 0É69). When used in production of raw shrimps, Oxinon is used at a concentration of 30È40 g litre~1, and BL-7P and Glaze-Nole are used together at concentrations of 15 g litre~1 and 10 g litre~1, respectively. Treatment with Glaze-Nole has a minor e†ect only, while treatment with BL-7P or Oxinon strongly reduces white spot formation (Table 1). A tendency to further reduction in white spot formation was observed when Oxinon concentration or the time of treatment was increased, however, these e†ects were not signiÐcant (Table 1). The e†ect of pH of the “chemical solutionÏ on white spot formation was studied for bu†er solutions with pH in the range 4É5 to 7É6 (Table 1, E†ect of pH). In general, pH decreased by 0È1 units during treatment of 6 kg shrimps, most for sea water. No direct e†ect of solution pH on white spot formation was observed. Phthalate bu†er treatment (pH 4É5È4É7) had no signiÐcant e†ect, while phosphate bu†er treatment (pH 5É4È 6É1) resulted in less white spot formation compared to sea water (pH 7É7) and borax bu†er treatment (pH 7É0È 7É6) increased white spot formation. The e†ect of sulphite, the major ingredient of BL-7P and Oxinon, was investigated for three concentrations (Table 1), and a strong inÑuence of concentration of sulphite on inhibition of white spot formation was observed. Sulphite concentration as in the usual BL-7P solution (0É032 M) has less inhibitory e†ect than BL-7P, whereas 0É048 M sulphite protects just as well as BL-7P. Sodium diphosphate, sodium tripolyphosphate and sodium trimetaphosphate in concentrations as in the usual BL-7P solution, with or without addition of phthalate bu†er to adjust pH to c 5É4, provide no signiÐcant protection from white spot formation (Table 1). The time it takes to pass a shrimp catch through the production process inÑuences white spot formation. Results in Table 3 show that shrimps that have waited for 1 h before chemical treatment and freezing (Time-2) W hite spots in the shell of raw shrimps during frozen storage 439 storage at [25¡C. The white spot formation in this case was as would be expected for shrimps stored at a constant temperature of c [27¡C. Progress in white spot formation in shrimps stored for six days at [16¡C and [18¡C, respectively, followed by storage at [25¡C, is well described by a combination of the rate at [16¡C or [18¡C and the rate at [25¡C. The temperature dependence of white spot formation is well described by an Arrhenius type equation (Fig 5) with an apparent energy of activation of 90(^8) kJ mol~1. The high energy of activation indicates that the rate-limiting step in the reactions leading to white spot formation is not enzymatically controlled and not di†usion controlled. The white spot formation is probably the results of a sequence of reactions, which may include enzymatic processes and di†usion of reactants. However, the rate limiting step is most likely a non-enzymatic hydrolysis. Fig 4. E†ect of storage temperature on development in white spot index. Observed index values are shown with symbols (L and …), and predicted index values are shown as full and dotted lines. A, storage at [16¡C (…, full line) and [25¡C (L, dotted line) ; B, storage at [18¡C (…, full line) and [29¡C (L, dotted line) ; C, storage at [16¡C for 6 days followed by storage at [25¡C (…, full line), and at alternating [25 and [29¡C (L, dotted line) ; D, storage at [18¡C for 6 days followed by storage at [25¡C (…, full line). have a higher degree of white spots than shrimps treated with chemicals immediately after catch (Time-1 and Time-1a). It had no signiÐcant e†ect if the shrimps were placed in the freezer immediately after packaging (compare Time-1a with Time-1). Storage temperature strongly a†ects white spot formation, and the rate of white spot formation increased with increasing temperature in the range [29¡C to [16¡C as can be seen from Fig 4. Alternating storage at [25¡C and [29¡C resulted in increased white spot formation compared to constant storage at [29¡C, and decreased white spot formation compared to constant Fig 5. Arrhenius type plot for white spot formation in the shell of pink shrimp during frozen storage. Rates are relative to the rate at [16¡C. DISCUSSION Like other crustacean exoskeletons, the shrimp shell consists of a thin outer epicuticle composed of protein, lipid and calcium salts, and a thicker procuticle composed of chitin, protein and calcium salts (Stevenson 1985). The cuticle, which also contains water, consists of several layers, and the white spots are formed within these layers. We have recently identiÐed the chemical composition of white spots as calcite and vaterite in close association with chitin (Mikkelsen et al 1997). Calcium and carbonate are natural components of the shrimp shell though no (visible) calcium carbonate crystals are formed in the living shrimp. From experiments where the development in white spot formation was followed over a time period (eg in Figs 2, 3 and 4), it is clear that white spots are initially formed as very small spots which grow in size, eventually covering the whole animal. This is in agreement with the identiÐcation of white spots as calcium carbonate, the precipitation of which starts as small crystals which grow in size. After catch, when the physical conditions as well as physiological conditions of the shrimp change, a number of factors might initiate crystallization and promote crystal growth in the shrimp shell. As a simple model, the aqueous phase of the shell can be considered as an aqueous solution of calcium and carbonate ions. When the shrimp is frozen, water forms ice, resulting in increased activity of solutes. Crystallization of calcium carbonate will occur when the solubility product constant, K , is exceeded sp K \ aCa2` É aCO2~ (7) sp 3 where aCa2` is activity of calcium ions and aCO2~ is 3 activity of carbonate ions. The solubility product constant, K , varies with temperature, salinity, crystal sp 440 form, purity and particle size (Simkiss 1976). Besides the precipitation equilibrium Ca2` ] CO2~ ¢ CaCO (s) (8) 3 3 the following gas/solution equilibria must also be taken into account CO (g) ¢ CO (aq) (9) 2 2 CO ] H O ¢ H CO (10) 2 2 2 3 H CO ¢ HCO~ ] H` (11) 2 3 3 HCO~ ¢ CO2~ ] H` (12) 3 3 From eqns (8) and (12), calcium carbonate precipitation would be followed by a fall in pH, as is also seen in laboratory experiments (eg Reddy and Nancollas 1973 ; Reddy 1975 ; Reddy and Wang 1980 ; Koutsoukos and Kontoyannis 1984). However, during storage of shrimps, pH increases as the result of other processes (Bhobe and Pai 1986 ; Riaz and Qadri 1990), the pH increasing more rapidly at increasing storage temperature (Riaz and Qadri 1990). White spots usually appear in the carapace (head shell) Ðrst, probably resulting from a larger increase in carapace pH than in somite (body shell) pH. During storage for 1 month at [25¡C, somite pH increased approximately 0É3 units from initially pH \ 6É9, while carapace pH increased approximately one unit from initially pH \ 7É9 (Royal Greenland, unpublished data). The observed increase in white spot formation with increasing storage temperature (Fig 4) is in agreement with the more rapid rise in pH at higher storage temperature. From the e†ect of pH of the chemical solutions on white spot formation (Table 1) it can be inferred that the chemical solution does not solely a†ect the reactions of calcium carbonate precipitation (eqns 8È12) as this would show an increase in white spot formation with increasing pH of the chemical solution. The pH of the chemical solution must a†ect white spot formation indirectly, eg via e†ect on hydrolytic reactions, protein conformation etc. Alternatively, the observed e†ect of the chemical solutions may arise from speciÐc inhibiting e†ects of compounds in the mixture on calcium carbonate crystallisation. In non-biological systems, calcium carbonate crystal formation and growth is inhibited by many organic and inorganic substances (eg Reddy and Nancollas 1973 ; Reddy 1975 ; Reddy and Nancollas 1976 ; Reddy 1977 ; Reddy and Wang 1980 ; Sohnel and Mullin 1982 ; Koutsoukos and Kontoyannis 1984 ; Meyer 1984 ; Sawada et al 1990). In the present study, retardation of calcium carbonate crystal growth by sulphite and phosphate was demonstrated in shrimp shell. E†ects of di-, tri- or trimetaphosphate at concentrations equal to their concentrations in commercial products for treatment of shellÐsh is, however, marginal. The addition of phosphates to commercial products for treatment of shellÐsh might have other advantages such as increasing water A Mikkelsen, B RÏnn, L H Skibsted binding. These compounds may, however, have a synergistic e†ect with other compounds on the inhibition of calcium carbonate crystallisation. A synergistic e†ect of compounds in Oxinon and BL-7P is indicated by a smaller e†ect of 0É032 M sulphite than that of BL-7P with the same concentration of sulphite (Table 1). The mechanism of inhibition of calcium carbonate growth is usually explained as a reduction in the reaction rate due to surface adsorption of the inhibitor at growth sites on the crystal surface (Reddy 1977 ; Morse 1990) with different efficiencies depending on di†erent adsorption affinities which result from di†erent charge, ionic radii and hydration properties of the inhibitor. Structural differences between SO2~ and CO2~ should be noted. In 3 3 the present investigation we have found that the length of time from catch to chemical treatment is critical for white spot formation. Changes in physical and physiological conditions when the shrimps are caught result in disturbance in the regulation of pH and ion concentrations in the shrimp shell, in e†ect initiating crystal seed formation and crystal growth of calcite and vaterite. To be e†ective, inhibition of calcium carbonate crystal growth should therefore be performed as early as possible in the production process, as clearly indicated from data in Table 4. The time from chemical treatment to freezing seems, however, not to be important for later formation of visible calcite. To accommodate the growing shrimp, the external skeleton must be shed and replaced periodically by the molt process (Gnatzy and Romer 1984 ; Stevenson 1985). The time of molting varies geographically and is inÑuenced by temperature (Shumway et al 1985), and might vary slightly from year to year. Before the old skeleton is shed, parts of it are digested and resorbed. A new cuticle is formed, followed by shedding of what is left of the old cuticle. Resorbed ions and other substances are used in formation of the new skeleton. Movement of ions during the period of molting might be a condition of easy calcite/vaterite crystal formation and crystal growth, as observed for two intervals during a period of 22 months (Fig 1). The data in Fig 1 can not be correlated to molt period, as this has not been recorded. However, in the geographical area where the present work was carried out shrimps molt mainly in MayÈJune. Another possible explanation of the increased incidence of white spots in the summer months might be the increased risk of periodically elevated temperatures during unloading, transportation, and storage in the warmest period of the year. The strong temperature dependence (Fig 5) of the rate limiting step in calcium carbonate crystallisation in shrimp shell means that careful control of storage temperature is an important tool in avoiding white spot formation. The high energy of activation (90 kJ mol~1) of formation of white spots indicates that the process of white spot formation is neither controlled by di†usion nor by an enzymatic reaction but more likely by a W hite spots in the shell of raw shrimps during frozen storage simple hydrolysis reaction. A very convincing e†ect of temperature on formation of white spots is also seen when frozen shrimps, without any white spots present, are thawed under warm tap water. Within a few minutes white spots appear in the shell. Therefore, the thawing process at inspection of the shrimps as well as by the consumer is very important to the observed incidence of white spots in the product. If shrimps without white spots are thawed under cold (O15¡C) water and the shell is dried at room temperature, no white spots appear upon heating to 100¡C, probably because calcium and carbonate ions are kept at a distance in the chitinÈprotein network of the shell during drying at room temperature. On the other hand, wet heating of shrimp shell facilitates movement of ions and promotes precipitation, an e†ect which is further enhanced by the solubility product of calcite decreasing with increasing temperature. CONCLUSIONS Formation of white spots in the shell of raw shrimps during frozen storage can be considered a natural phenomenon, as white spots always appear after some time of storage at [25¡C, if the shrimps are not treated chemically prior to freezing. Chemical treatment with sulphite, phthalate and phosphates was found to decrease calcite crystallisation in shrimp shell, while treatment with borax promotes crystallisation. Increased time between catch and chemical treatment as well as increased storage temperature results in increased formation of white spots. As several factors are shown to inÑuence white spot formation, control of a single factor might not result in a satisfactory improvement in the visual quality. In contrast, the simultaneous presence of several factors such as long production time and low concentration of crystal inhibitors in the chemical bath, followed by high temperature during unloading, transport and storage, will certainly cause substantial reduction in the shrimp quality. An increased risk of white spot formation is observed in the summer period. This might result from changes in ion regulation in the shell during molt, a factor which should be considered in future investigations. ACKNOWLEDGEMENTS This research was a part of the F^TEK programme sponsored by the Danish Ministry of Research and the Danish Ministry of Agriculture and Fisheries and conducted as a collaborative project with the company Royal Greenland (Roslev, Denmark) and LMC-Centre 441 for Advanced Food Studies. Arctic Station (University of Copenhagen) in Qeqertarsuaq, Greenland, is thanked for providing laboratory facilities. REFERENCES Bhobe A M, Pai J S 1986 Study of the properties of frozen shrimps. J Food Sci T echnol 23 143È147. Gnatzy W, Romer F 1984 Cuticle : formation, molting and control. In : Biology of the integument. 1. Invertebrates, eds Bereiter-Hahn J, Matoltsy A G & Richards K S. Springer Verlag, Berlin, Germany, pp 638È684. Koutsoukos P G, Kontoyannis C G 1984 Prevention and inhibition of calcium carbonate scale. J Crystal Growth 69 367È376. McCullagh P, Nelder J A 1989 Generalized L inear Models. Chapman and Hall, London, UK, pp 99È135. Meyer H 1984 The inÑuence of impurities on the growth rate of calcite. J Crystal Growth 66 639È646. Mikkelsen A, Engelsen S B, Hansen H C B, Larsen O, Skibsted L 1997 Calcium carbonate crystallization in the a-chitin matrix of the shell of pink shrimp, Pandalus borealis, during frozen storage. J Crystal Growth 177 125È134. Morse J W 1990 The kinetics of calcium carbonate dissolution and precipitation. In : Reviews in Mineralogy (Vol 11, Carbonates : Mineralogy and Chemistry), ed Reeder R J. BookCrafters, Michigan, USA, pp 227È264. Reddy M M 1975 Kinetics of calcium carbonate formation. V erh Internat V erein L imnol 19 429È438. Reddy M M 1977 Crystallization of calcium carbonate in the presence of trace concentrations of phosphorus-containing anions I. Inhibition by phosphate and glycerophosphate ions at pH 8É8 and 25¡C. J Crystal Growth 41 287È295. Reddy M M, Nancollas G H 1973 Calcite crystal growth inhibition by phosphonates. Desalination 12 61È73. Reddy M M, Nancollas G H 1976 The crystallization of calcium carbonate IV. The e†ect of magnesium, strontium and sulfate ions. J Crystal Growth 35 33È38. Reddy M M, Wang K K 1980 Crystallization of calcium carbonate in the presence of metal ions. I. Inhibition by magnesium ion at pH 8É8 and 25¡C. J Crystal Growth 50 470È480. Riaz M, Qadri R B 1990 Time-temperature tolerance of frozen shrimp 2. Biochemical and microbiological changes during storage of frozen glazed shrimps. T rop Sci 30 343È356. Sawada K, Ogino T, Suzuki T 1990 The distribution coefficients of Mg2` ion between CaCO polymorphs and solu3 and transformation of tion and the e†ects on the formation CaCO in water. J Crystal Growth 106 393È399. Shumway3 S E, Perkins H C, Schick D F, Stickney A P 1985 Synopsis of Biological Data on the Pink Shrimp, Pandalus borealis KrÏyer, 1838. FAO Fisheries Synopsis No. 144, NOAA Technical Report NMFS 30. Simkiss K 1976 Cellular aspects of calciÐcation. In : T he Mechanisms of Mineralisation in the Invertebrates and Plants, ed Watabe N & Wilbur K M. University of South Carolina Press, Columbia, pp 1È31. Stevenson J R 1985 Dynamics of the integument. In : T he Biology of Crustacea (Vol 9 Integument, Pigments and Hormonal Processes), eds Bliss D E & Mantel L H. Academic Press, New York, USA, pp 1È42. Sohnel O, Mullin J W 1982 Precipitation of calcium carbonate. J Crystal Growth 60 239È250.