The data shown in Table 1, Appendix 2, are a summary of the insolation, air speed, ambient temperature and humidity, and internal solar dryer temperatures as recorded over the experimental period.
Figures 4 and 5, Appendix 2, illustrate the hourly variation in incident insolation and internal temperature for the three solar dryers over the course of a day. Figure 4 is typical of a fine day with high insolation and Figure 5 of a cloudy day with low insolation. It can be seen that for the former with insolation of 17,636kJ m2 over the 9-hour period, the internal temperatures of the tent dryer, SCD dryer and cabinet dryer averaged 18°C, 19°C and 19°C higher than ambient, whereas for the cloudy day with insolation of only 8,876kJ m2, their internal temperatures averaged 14°C, 11°C and 14°C higher than ambient. Although, as can be seen from Figure 5, the tent and cabinet dryers achieved higher internal temperatures than did the SCD dryer, examination of Table 1 shows that no one dryer consistently attained higher average daily internal temperatures. It must be borne in mind, however, that the vents and doors of the dryers were more fully opened in the period around midday on sunny days, in order to prevent fish from being cooked or case-hardening occurring; had this not been done then the difference in temperature elevation between fine and cloudy days would have been more marked.
On occasions, the temperature distribution within the solar dryers was tested and it was found that variation of temperature in the vicinity of the drying rack (or trays) was very small: a spread of 4°C was the maximum encountered.
Drying curves obtained for the three solar dryers and the two sun-drying methods for the experiments completed are shown in Figures 6 to 13, Appendix 3, and condensed data showing the drying times for specific changes in moisture content are presented in Tables 2 to 9, Appendix 3. Data for Experiments 5 and 7 are not included because these were prematurely terminated since the drying rates were not sufficiently high to prevent putrefaction of the fish within a few hours of it being set to dry. From these curves and Tables, it is obvious that the solar dryers gave higher drying rates overall than the sun-drying methods; also the solar-dried fish had lower final moisture contents than those dried in the sun, which will be discussed in section 3.
Upon close examination of the data, it can be seen that the difference in drying rates between solar and sun drying was relatively small in the initial stages; indeed for Experiments 1 and 4, as shown in Tables 2 and 5, Appendix 3, the times required to dry down to 40% moisture were very similar for both sun and solar drying. However, as drying proceeded, the rate of moisture loss for the sun-drying methods decreased at a higher rate than did that of the solar dryers. Of the latter, the tent and SCD dryers performed in a very similar manner throughout but the performance of the cabinet dryer was intermediate between the solar dryers and the sun-drying methods, except at low moisture contents (< 30%), when its drying rate was fastest of all, thus enabling it to catch up on the tent and SCD dryers.
Averaged data showing the relative performance of each dryer and sun-drying method for certain changes in moisture content are presented in Table 10, Appendix 3, while Figure 14, Appendix 3, illustrates generalised drying curves for each dryer and sundrying method. Consideration of the controlling parameters of the drying process, air flow and air temperature, and an understanding of the operating principles of solar dryers can help to provide an explanation for the observed dryer performances as outlined.
In the initial stages of drying, it is usually the case that the surface of the material is covered with a thin film of moisture which is continually being removed by the flow of air over the surface and replenished at an equivalent rate by the migration of moisture to the surface from the interior of the material. The rate of drying therefore is very dependent upon the rate of air movement across the surface. The temperature of the air is of much less importance provided that it is above the dewpoint and that sufficient heat is available to provide for the latent heat of evaporation.
In the final stages of drying, the surface of the material can be regarded as effectively free of moisture, with all of that remaining contained within the interior. In this situation the rate of drying is controlled by the temperature gradient from the interior to the surface which in turn is controlled by the temperature of the surrounding air or by radiation incident upon the surface such as insolation. As the rate of moisture removal is considerably less than in the initial stages of drying, the degree of air movement across the surface is of lesser importance. In brief, drying rate is controlled by air movement initially, and by air temperature in the final stages, with both exerting an influence in the intermediate stages.
Relating this to the results obtained, the similarity in drying rate between solar and sun-drying techniques in the initial stages can be explained by the hypothesis that the air movement provided by the natural breezes around the fish spread on the rocks and rack was as effective as the flow of air within the solar dryers caused by natural convection, and that the higher temperatures within the latter had little significant effect. As the fish dried, the degree of air movement exerted progressively less influence, and the higher temperatures within the solar dryers progressively more influence, resulting in the higher drying rates during the later stages.
The difference in performance of the cabinet dryer compared with the tent and SCD dryers may be explained by consideration of the mechanisms operating within the cabinet dryer. The rate of air movement caused by natural convection is proportional to the height of the 'air column' over which the temperature gradient exists. As this height is considerably less for the cabinet dryer than for the tent or SCD dryer, and similar temperatures were attained within each solar dryer, then it is reasonable to hypothesise that the flow of air through the cabinet dryer would be correspondingly less than through the others.
In Figure 14, Appendix 3, the curve BC can be explained thus: in this intermediate stage of drying where both air movement and air temperature exert an influence, the drying rate of the cabinet is less than that of the tent and SCD dryers because of the smaller air flow, but greater than sun drying because of the higher temperature within the cabinet. Within this dryer the fish are placed much closer to the 'hot' black surfaces than in the others and it is reasonable to suppose that the fish themselves would, therefore, attain a higher surface temperature. Unfortunately, it was not possible to verify this experimentally. In Figure 14, the curve CD lies in the region where one would expect temperature to be the rate-controlling parameter, and therefore the higher fish temperature would account for the higher drying rate.
From Table 10, Appendix 3, it can be seen that to dry to a final moisture content of 20%, an acceptable figure for dry salted fish, the solar dryers required 60 - 65% of the time necessary for sun drying: 3 days compared with 5 days. These data enable an estimate of production rates to be made. Under sunny conditions, it can be said that solar-dried fish would require 3 days to attain 20% moisture. Given a capacity for the solar dryers of 8 kg of prepared fish per square metre and a 50% weight loss with dry salted products, then the rate of production of dried salted fish would be approximately 1.25 kg m-2 day-1 . Similar calculations for sun-drying give a production rate of approximately 0.75 kg m-2 day-1 . It should be remembered, however, that these are only indicative figures since they will be affected by shape, size and species of the fish, the nature of the final product (e.g. moisture and salt contents, fillets, split, etc.) and weather conditions.
The difference between the performance of the two sun-drying methods is slight but interesting; Table 10 shows that the fish on the rack initially dried at a faster rate but, at lower moisture contents, their drying rates were lower than those on the rocks. A possible explanation for this might be that the initial higher rate could be due to the better air circulation around the rack, whereas in the final stages, the faster drying of fish spread on the rocks could be due to the higher local air temperatures surrounding the fish as a direct result of their proximity to the black lava rocks.
When purchased at the landing site, the lisa and pampano were of excellent quality: they had clear, convex eyes, bright red gills, strong fresh gill and body odours, bright and glossy skins, a firm texture, and several were in rigor mortis. Since ice was not available, they were held in seawater in a shaded position and were processed as soon as possible, thus ensuring only high-quality fish were used for the drying experiments.
Visual examination of the final dried salted products indicated that the solar-dried fish were of good quality and were marketable. The texture was hard and well dried and the products had a pleasant odour. The dry salted fish were of a light yellow colour, whilst the brined products were dark orange, showing signs of rancidity. Sundried products were of poorer quality, particularly the fish dried on the lava rocks; there was sand and dirt adhering to the flesh and the fish had suffered from attack by insects, birds and other animals. However, the-sun-dried products were of better quality than locally-produced dried salted fish, which were generally well salted, but insufficiently dried and had a very strong unpleasant odour. None of the experimental products showed any evidence of mould attack, 'pinking' caused by salt-tolerant bacteria, beetle infestation, or case-hardening.
Under the conditions of these experiments, regardless of the drying method employed, it was not possible to reduce the moisture content of unsalted fish fast enough to retard spoilage. By the second day of drying, these unsalted fish were putrid and, therefore, the trials were discontinued.
It is known (Scott, 1957) that the stability of salted and dried food products depends on their water activity (aw). This is a measure of the free or available water in a food which is able to react chemically or, in spoilage, to support the growth of microorganisms, such as bacteria and moulds (Waterman, 1976). The relationship between moisture content and aw is usually expressed in the form of a sorption isotherm. For dried salted fish it is necessary to consider both the moisture content and the salt content when calculating aw (Doe et al., 1982). Relatively high concentrations of salt have an inhibitive effect on micro-organisms, an obvious exception being the red halophilic bacteria which are salt-tolerant. The aw of pure water is assigned the value of 1 and the aw of a food is expressed as a fraction relative to pure water. Fresh fish have an aw of above 0.95. Most spoilage bacteria will cease to grow in a food whose aw is below 0.90 and the growth of most moulds is inhibited below 0.80. However, halophilic bacteria can grow at an aw as low as 0.75 and some xerophilic moulds as low as 0.65 (Bone, 1969).
Salting and drying both have the effect of reducing aw. During storage, dried fish flesh will absorb moisture from the air at high humidities until an equilibrium is reached. At humidities of above 75%, any salt in the flesh will also absorb moisture. Therefore, during storage, the product becomes wetter, thus increasing the aw, and it will become more susceptible to spoilage by moulds and bacteria. Fish processed in hot, humid tropical climates by salting and subsequent drying are liable to deteriorate during storage through mould growth; those xerophilic moulds causing dun and the brown discoloration associated with localised decay are the most troublesome (Liston, 1980). Poulter et al. (1982) have used isohalic sorption isotherms for cod combined with growth data for the dun mould, Wallemia sebi, to predict the mouldfree shelf-life of dried salted fish under tropical conditions.
The results of the moisture and salt analyses of the final dried salted products, expressed on a wet weight basis are given in Table 11, Appendix 4. They indicate that lower final moisture contents were achieved in the solar dryers: the average moisture content of solar-dried fish was approximately 13%, while that of sun-dried fish was approximately 21%. Dry salted products had an average salt content of 25% and the brined products of 13%.
Using these moisture and salt contents and the method of calculating aw and estimating shelf-life outlined by Poulter et al. (1982), it can be shown that all the solar dried products had an aw of below 0.65 and a predicted mould-free shelf-life of over 450 days. The sun-dried products had an aw of 0.65 and above and would have a predicted mould-free shelf-life of between 100 and 450 days. However, since lisa belong to a family which is of medium fat content and pampano to a family of high fat content (Sidwell et al., 1974), the actual shelf-life of the dried fish may be shorter due to the inevitable onset of rancidity.
During these experiments, the fish were dried until there was no further appreciable weight loss. In commercial practice, it might not be necessary to dry the products to the same extent. Again using the method of Poulter et al., it is possible to suggest that, with a moisture content of 25% in the brined products, or 25 - 45% in the dry salted products (aw = 0.70), a shelf-life of approximately 100 days might be expected. If a longer shelf-life is required, the moisture content would need to be reduced to 15 - 20%, regardless of the salt concentration, to obtain a more stable product. As mentioned earlier in this report, it was not always possible to achieve this lower moisture level during sun drying. Therefore, when using the solar dryers, it is not only possible to achieve higher drying rates but also lower moisture contents and, in turn, a more stable product.
One point to remember, however, is that the actual rate of drying, and therefore the rate of reducing the aw, will affect product quality. Spoilage will occur during drying and will not be retarded until the aw is sufficiently lowered. Therefore, due to the different drying rates, products dried in the cabinet would probably have deteriorated more than products dried in the other two solar dryers, but less than the sun-dried products. This would not necessarily be apparent at the completion of drying but would become so during storage.
Acid-insoluble ash is a useful index of mineral matter, such as dirt or sand, in foodstuffs (Pomeranz and Meloan, 1978). The results of acid-insoluble ash analysis on the final dried salted products, calculated on a dry weight basis, are given in Table 11, Appendix 4. On average, those products which were dried in the solar dryers and on the rack had 1.0% or less acid-insoluble ash whilst those dried on the rocks had 1.3%. These results confirm two expectations: firstly, the fish dried on the rocks were generally more contaminated by dirt, etc. than any of the other fish and, secondly, it was possible to obtain as clean a product by drying on the rack as it was by solar drying.
During the drying process, and subsequent storage of the dried fish, insect infestation is a major problem. In addition to causing losses in quality and quantity, insect pests are potential carriers of pathogenic bacteria and thus represent a serious health hazard (Proctor, 1977; Wood, 1981).
Flies, the major carriers of disease, lay eggs on fish during the early stages of drying, becoming less attracted to them as the flesh dries and hardens. The larvae tunnel into the flesh, causing putrefaction and extensive physical damage. The most important pests of the dried fish are beetles of the family Dermestidae They will invade the fish flesh from the earliest stages of drying but, unlike flies, will continue to be attracted to, and breed in, the dried product.
Blowflies were a large problem during drying in this exercise, particularly with the fish dried on the rocks which were heavily infested with eggs and larvae; infestation also occurred in the rack-dried fish but to a lesser extent. The solar-dried fish showed little evidence of blowfly attack since the flies were killed by the high temperatures within the solar dryers.
When fish are salted, the primary effect of the salt is bactericidal but it also retards insect infestation (Proctor, 1977). This is supported by the evidence of these experiments where blowflies were less attracted to the dry salted products compared to the brined products, but during sun drying the unsalted fish were usually completely covered by flies. When drying brined or unsalted fish, it was necessary to wash the rocks to remove the large piles of eggs laid by the flies.
When loading the tent, the fish were carried into the dryer and spread on the drying rack. However, with the cabinet and SCD dryers, loading was accomplished from outside and a large number of flies were attracted by the fish and, subsequently, ventured inside the dryers; this was particularly noticeable with the cabinet. Escape from the cabinet was more difficult for the flies because the doors were closed immediately following loading. On days of high insolation, most of the flies in the cabinet were killed within three hours of loading the dryer each morning, by which time the internal temperature had reached 55°C or above. When the flies were killed by the high temperatures, they fell onto the fish. This can result in faster spoilage of the product before drying is completed (especially at the higher temperatures achieved within the dryer) and could also be a potential health hazard: both of these problems are due to contamination of the fish by the bacteria carried by the flies. As a result, it is not possible to recommend the use of the cabinet dryer for fish unless further steps are taken to prevent entry of flies. The use of mosquito netting, or a similar material, draped over the doors whilst they are partially open in order to reduce the entry of flies, may help to lessen the problem, but would increase operating costs and involve a more complex operating procedure.
Most of the dead flies in the tent and SCD dryers were found on the floor. The flies are easily removed by sweeping the tent floor, but some means of entry into the SCD dryer is necessary for cleaning purposes. This could be accomplished simply by positioning a flap below the drying rack.
No evidence of beetle attack was found on any of the batches of fish at the end of the drying period. There were dermestid beetles found around the storage area, however, and two months after processing a few dead larvae were found on samples of brined lisa.
A point worthy of mention at this stage is the possible use of solar drying as a means of de-infesting sun-dried fish; the elevated temperatures would kill any insects or lava present on the fish. Szabo (1970) has carried out some investigations on this topic with promising results.
Material costs for the solar dryers and sun drying methods were estimated as detailed in Appendix 5 and are summarised as follows:
It must be emphasised that these costs are specific to the Galapagos and should not be regarded as applicable to the mainland of Ecuador or elsewhere. Some items, for example plastic sheet and plywood, are particularly expensive in the Galapagos because of the high costs involved in shipping these materials from the mainland.
These costs indicate that there is little difference in material costs between the tent and SCD dryers and that the cabinet is appreciably (c. 40%) more expensive. When construction times are also taken into account (6, 15 and 25 hours respectively for the solar tent, the SCD dryer and the solar cabinet), then the tent dryer appears the simplest and cheapest to construct. Obviously if different, preferably cheaper, materials were used for the solar dryers, then the costs would be altered. Substitute materials which would reduce the cost might be: locally-available coarsely woven material or basket work as an alternative for the plastic mesh for the racks and trays, and walls of local brick or mud construction for the cabinet instead of plywood.
Although in the short time available it was not possible to obtain figures, it is reasonable to assume that the major component in operating costs, excluding that of labour, would be that of maintenance, and in particular that of replacing the plastic sheet. A high proportion of the material cost of the tent and SCD dryers is that of the plastic sheet, 43% and 39% respectively, compared with only 2% for the cabinet dryer. It can therefore be expected that the operating costs of the former will be considerably higher than that of the cabinet.
It has been suggested that clear polyethylene sheet that has been rendered more resistant to degradation by the ultra-violet component of sunlight in order to improve its effective lifetime would reduce operating costs. However, it would seem more likely that physical wear and tear would be the major factor limiting the lifetime of the plastic.
For the purposes of this exercise the solar dryers were used in their recommended manner with only one rack in the drying chamber. However for the tent and SCD dryers, it is considered that production rates could probably be increased and production costs decreased either by the use of additional racks or by hanging the fish within the dryers rather than spreading them on racks.
The dryers used for this exercise were sized for experimental purposes but for some commercial operations there would be a need for larger and more cost-effective dryers. For such dryers, increased attention would have to be paid to control of air flow and temperature, structural stability, especially in areas of high winds, and ease of loading.