A major problem at Aiyura appears to be the poor identification of priorities. The decision has been made to produce table-size carp in ponds with ducks and to grow vegetables on the banks. The Aiyura site is expected to grow a total of 6 tons of produce in 1986. Of this 4 tons are to be carp, produced in two half-yearly harvests each from two ponds with total area a little over 0.4 ha (ponds 2 and 4). The present plan is that each 2 000 m2 “integrated” pond is to have 30 ducks living nearby and swimming on it during the day. As a fully-grown laying duck produces about 150 g of excreta per day on a dry weight basis, the quantity of manure will start negligibly small when the ducklings are newly-hatched and gradually increase to a total of not more than about 4½ kg/day. This maximum quantity is about one tenth of what is required. (Edwards, 1985, used 30 full-grown laying ducks on 200 m2 ponds in Thai villages). So fish growth based on the integrated farming concept, i.e., that the ducks are fed and the fish are not, will only occur while the fish are very small; thereafter they will have to be fed. The projected level of fish production from these two large ponds with a few ducks on each is equivalent to 8 – 10 tons/ha/year. The consultant found this unrealistically high and suggested that the production from modestly-manured ponds may reach 1–2 tons/ha/year. Any higher production would require high inputs, polyculture approach and very high competence of management.
During the consultancy, i.e., between November 1985 and February 1986, the Aiyura site, which was being maintained without production, was reactivated. This included rehabilitation of some of the existing ponds; the broodstock was doubled in size; reasonable numbers of fry were cultured and growing trials started in specially-constructed small experimental ponds. However there is still great need for trials on coffee pulp and other agricultural wastes (see Fish Feed Specialist's report) to be commenced, for extension to begin again and for local people to be recruited into cooperative experimental work on pond building, stocking, feeding, manuring and harvesting systems. At present, if preference is being given by the DPI to the “integrated” farming experiment taking most of the pond space for the next year or more, such activities cannot be expected to take place for some time to come.
If progress is to be achieved the following aspects should be given priority:
Breed carp and distribute large fingerlings (not fry).
Test carp growing systems using locally-available wastes and byproducts.
Cooperate closely with selected local farmers to set up pilot schemes with monitoring of pond inputs and yields.
Dig more small ponds for experiments on site using existing spare land and labour.
Carry out growth trials of fish in small unfed ponds manured by fed poulty.
Gain experience in polyculture, initially using locally available species.
Test the use of animals on site to reduce the need for grass cutting and provide manure.
Extend fish culture more widely demonstrating only tested growing methods and insisting on information feedback in exchange for fingerlings.
Carp breeding and growing systems, polyculture, raising fish with poultry and aspects of site improvement and maintenance are dealt with in more detail below.
Until the Aiyura site is extended (in accordance with the plans of aquaculture engineer J. Kovari, a member of the present team), there are only four large ponds available. Depending on the priority given to different tasks it is likely that four, three or only two of these large ponds, plus several small holding ponds, will be available for carp breeding and fingerling production. These alternative cases are discussed below.
(i) Four-pond breeding plan
If it is decided to do the preliminary experiments on duck plus fish farming in small ponds so that four large ponds are available for producing fish for distribution then the work could be planned as follows. About half the present stocks of large females and males should be put into pond l. This would come to around a dozen of each. In addition 20 – 40 of the largest fingerlings from pond 4 which are not obviously males should also be placed into pond l; these are to be grown until they form an additional group of broodstock. The pond should then be fertilized and manured. There should be frequent partial croppings, using a net after lowering the water level through a fine screen.
The other half of the female broodstock should be held in one of the remaining large ponds, either pond 2, 3 or 4, taking care that males get no chance to enter. With pond manuring (and possibly feeding) each of these fish should become ripe for spawning within 3–4 months of a previous spawning, if held at low density. This should allow 2–4 fish to be spawned each month. Therefore, once or twice a month the broodstock pond should be emptied and all the fish checked. Ripe fish should be spawned by one of the methods discussed in Appendix 3. Fry should be reared in another large pond. The fourth large pond can be used for holding males and for growing fry to fingerling size.
Ponds for fry rearing should be emptied before spawning is arranged, so that they can be fertilized about a week before larvae are introduced. Therefore it may be convenient to transfer the unripe broodstock from one large pond to another when checking them, and to put the fish selected for spawning into a small pond for a few days. Then prepare the newlyemptied broodstock pond for the coming fry rearing before carrying out spawning. Ripe broodstock may be held in small ponds for some time without loss of ripeness but they are not likely to become ripe when held at high density. Some males can be held in small ponds and should remain ripe.
If space for fry or fingerling rearing becomes limited the following system may be tried. After checking broodstock put ripe ones into one small pond and unripe ones into another. Prepare the fry pond, spawn the fish and stock the fry pond. Put the spent females back into a small pond. After the fry have been in the pond for about 10 days, take the females from the small ponds in which they were being held and put them into the fry pond. Increase the feeding rate correspondingly. After another 3–4 weeks harvest the fry and carry out broodstock selection at the same time. Broodstock may also be ripened in fingerling ponds as long as there are only young fingerlings there (and not small adults as were found in pond 4). Systems of feeding broodstock at different densities are discussed in Appendix 3.
(ii) Three pond breeding plan
If three of the four large ponds are available for fry and fingerling production, it is recommended to use pond 1 for uncontrolled breeding and growing candidate broodstock. Additional fry can be stocked in it for raising to fingerling size. Males would have to be held mainly in small ponds, with a reserve stock in the pond used for producing table-size fish. The remaining two ponds would be for broodstock ripening and fry production. Use would have to be made of the various small ponds for holding broodstock for short periods, and broodstock ripened sometimes together with fry which have been in ponds for more than about 10 days or with fingerlings.
(iii) Two pond breeding plan
If only two ponds are available for breeding then this work will have to be on a reduced scale. It will become difficult to rear fry to fingerling size in adequate numbers and broodstock will have to be ripened in fry rearing ponds for much of the time.
Carp culture in the highlands has not been successful up to now. Quite a lot of villagers have made small ponds and put a few fish inside. A number of farmers or institutions have built ponds of a few hundred square meters or more. There are virtually no records on inputs and yields. There are clearly some places where carp are breeding in ponds. However, the overall contribution to the economy or nutrition of the highlands has been close to zero.
The explanations for this vary according to the attitude of the people making them. The lowland fisheries people generally seem to believe that highlands fish farming is totally unsuitable and unrealistic, a complete waste of effort and resources. The aquaculture enthusiasts think it is a wonderful idea but are currently saying that it has so far failed because of lack of extension work. The consultant is taking a middle view. There are indeed obstacles, such as the uncertain or communal nature of land ownership, insecurity (rascals, feuds), the lack of familiarity with animal husbandry (apart from free-range pig rearing), reluctance to handle manures, the high cultural value of pigs which get most of the left-overs, abundant alternative income sources (in some areas) and low cost of tinned fish.
On the other hand there is plenty of swampy land and lots of water during most of the year. People are interested; and new crops (coffee, timber, vegetables, chickens) have been introduced and taken up enthusiastically in many cases when they have proved worthwhile. This is the main point in the consultant's opinion. In all the years during which fish culture has been talked about, nobody, it seems, has done any experiments to see whether it is possible to manage fish ponds in a way which is profitable or which contributes to better nutrition. It has been proved that fish can be grown in the PNG highlands. But it has not been proved that they can be grown in a way which brings real benefit to the grower. Nobody seems to have looked at the inputs and outputs of ponds to see whether it is worth the effort. An attempt to do so is presented in Appendix 4, but it is all based on data from work done abroad. There are no data from here because the work hasn't been done. Until such information is gathered, failures cannot be blamed on the villagers.
Experiments are required to investigate stocking densities and manuring rates in fed and unfed ponds. Many short experiments in small ponds are likely to yield more information than a few long-running trials in big ponds. Fish must be weighed regularly (every 2–4 weeks) Systems tried should be those which villagers are able to afford using locally-available cheap materials.
The weight of fish per square metre is much more important than the number of fish per square metre. For fed ponds aim for fish biomass in the range 200–2 000 kg/ha (20–200 g/m2) over the growth period. For fertilized unfed ponds densities should be about half this. Stock fish of a range of sizes at one time or break the growth into several phases and thin the fish every few months rather than trying to carry out growout from fry to table size in one pond at a single stocking rate. Compare growth rates with those given in Table 2 (Appendix 4, section 2.4) and lower the stock rate if growth rates drop greatly.
Food can be given by hand according to the response of the fish once or twice a day. It can be given by demand feeder, allowing the fish to eat as much as they want. Or fish can be fed, with either of these two systems, but only giving them a pre-determined amount each day. This amount is usually calculated on the basis of periodic weighings as a percentage of the total weight of the fish in the pond. For feeding rates see Tables 1, 3 and 4 in Appendix 4, section 2.4 However, it must be remembered that these tables apply to dry feeds which are more or less completely digestible. If a material containing 75 percent water is used, one has to give fourt times as much. If a feed largely made up to vegetable fibres or husks is used, of which carp will digest very little, a similar adjustment will have to be made.
Systems of manuring and fertilizing ponds are discussed in detail in Appendix 3, section 5.2 and Appendix 4, section 5.3 Due to their high price in PNG it seems unlikely that inorganic fertilizers will be used by villagers in growing ponds. Materials which may prove useful and should be tested include animal manures, from poultry, pigs or cattle, and agricultural or gardening wastes, in particular coffee pulp. Composts produced from grass, weeds or vines plus animal manures may also be helpful. Moderate manuring rates involve the application of 50–200 kg of dry material per hectare per week while intensively-manured fishponds may receive over 1 200 kg/ha/week. This wide range of application rates of various materials will have to be tested.
Design of experiments
Preliminary trials can be carried out in very small ponds (i.e., 10 m2), but promising results should be tested again in ponds which are similar to those of villagers (50–200 m2). Growing systems in which fish of a range of sizes are present in the pond at one time utilize pond productivity more efficiently (see Appendix 4, section 5). In small-scale experiments this type of stocking has additional advantages. Choose fish which are well spaced in size and of different colours, and write a detailed description of each (weight, shape, colour, sex if known, and any unusual markings). Then when the fish are re-weighed, they can be identified and individual growth rates calculated. This protects the experiment against the loss of information which would otherwise occur should a few fish die. It also allows weighings in which not all the fish are caught to yield accurate data. This is important particularly when testing manured systems as otherwise ponds would have to be emptied frequently, with the loss of the fertilized water. Details of an example of this type of trial are given in Appendix 5, Note that negative results are as important as positive ones, and that a large number of fairly short experiments will be more instructive than a few long ones. Clear and complete recording of all data is essential. It is strongly recommeded that information be recorded directly into a field notebook as the work is carried out, not written on scraps of paper for later copying into a notebook which remains in the office.
Polyculture of several species of fish together in one pond offers the possibility of increased productivity by utilizing the various sources of food more efficiently. Of particular potential value are Chinese carps, Indian major carps and some tilapia species. The Chinese carps are very unlikely to breed without induction in local ponds or rivers. They therefore require a higher level of hatchery technique for propagation. However, they are also much less likely to threaten indigenous fish species when they get into local waterways.
In the future it is possible that the Government of Papua New Guinea may approve the introduction of exotic species either for aquaculture or for improving river fisheries. In that case polyculture possibilities will be greatly extended. In the meantime, however, there are a number of fish species present in the country which merit consideration and preliminary experimental work.
Oreochromis mossambicus is at present the only tilapia species in PNG. The only accounts from PNG dismiss this fish, saying that they do not grow to a good size in highland ponds, without giving details. This probably reflects on the management of the ponds in question. There are many ways to prevent or reduce the unwanted breeding of tilapia. These include growing them in cages, introducing a predator, monosex culture by hybridisation, monosex culture by sex reversalusing hormones and monosex culture by hand sexing. Of these the last is most likely to be useful for small ponds here. Tilapia which are more than 15–30 g are usually easily sexed to a fair degree of accuracy. Males grow faster; females should be discarded (deep-fried or baked over a fire perhaps) except for a certain number for breeding. A few mistakes do not matter too much as long as one can achieve 80–90 percent males. Polyculture in various proportions with carp can then be tried at different stocking densities, feeding and manuring rates.
To breed tilapia for sexing and stocking in polyculture the following system may be used. Two types of ponds are required: breeding ponds and fingerling ponds. The area of fingerling ponds will have to be at least ten times that of the breeding ponds. Both should have essentially static water of at least 60 cm depth and be well manured.
The largest available fish are stocked in the breeding pond at a ratio of about 3 females to each male and a density of up to one fish per square metre. When small fry are seen in the pond all the fish are removed by a combination of netting and draining the pond. Fry are transferred to the fingerling ponds where they are stocked at up to 5/m2 (carp fingerlings may be raised in the same pond as long as the stocking density is not excessive). Breeders are returned to the breeding pond for another cycle.
The recommended frequency for cropping fry from the breeding pool is once every 1–1½ months. The mouthbreeding stage of most eggs or yolksac larvae should be finished and a second batch not yet spawned. A set of breeders may be able to produce half a dozen batches of fry before their reproductive rate drops, when they should be replaced with fresh, wellnourished fish. Females of above 400 g can produce around 1000 fry at a time, but the small fish at present available at Ayiura are not likely to produce more than 50–200.
Grey mullets are marine fish, some of which can be grown in freshwater ponds. Mugil cephalus is the most important cultured species, grown in Italy, Israel and many parts of Asia. Woynarovich and Horvath (1980) mention M. tade and M. dussumieri as being grown in India, Pakistan and Indonesia. M. capito and a few M. auratus are also grown in Israeli polyculture ponds with M. cephalus, tilapia, common carp and Chinese carps.
Mullets have been spawned and larvae reared in captivity. However, this is difficult and yields have been very low. All practical culture is based on wild-caught fry. In Israel they are caught from December to February overnight or in the early morning following rain, when they are attracted into the mouths of rivers or streams where they meet the sea. The fry are then about 1½–2 cm in length and are extremely delicate. They must be caught with hand-nets or seine nets of fine fly netting or curtain cloth (pore size 1–2 mm) and transferred immediately to containers of good water. They are transported in tanks with aeration. The seawater is gradually diluted over a few hours or a couple of days, taking care that the water quality remains good and that they are not subjected to rapid temperature changes to which they are sensitive. Large plastic bags with about 20 litres of water and 40 litres of oxygen would probably be suitable for transporting a few thousand fry provided the water was clean and the temperature could be kept steady. Malachite green can help prevent fungal infections.
For experimental purposes larger mullet fingerlings can sometimes be caught by beach seining although many will jump over the net and escape. This may save up to a year of growth, though of course only small numbers are likely to be caught and transportation becomes more costly. Identification of fry is difficult, requiring dissection and microscopic examination.
Mullets are able to exploit detritus and films of organic material on the top of the water and on the surface of the mud. They may also scrape attached algae off hard surfaces and filter feed. Although rather more sensitive than carp to low oxygen levels, and very sensitive to poor handling, they are generally hardy and grow well under a wide range of conditions and in waters of high fertility without supplementary feeding.
Like mullets, Chanos chanos is a marine species the fry of which can be collected close to the shore. They also adapt to freshwater and a wide range of temperatures, surviving down to about 12°C. They exploit the products of fertilization, in particular microfilamentous blue-green algae growing on the mud surface, decaying filamentous green algae, and perhaps planktonic algae too. They are grown in polyculture, often with shrimp and tilapia, and are among the most important cultivated fish in SE Asia.
(iv) Sea bass (barramundi)
Lates calcarifer might be a very effective “police fish” to prevent unwanted breeding by tilapia in ponds. They are said to be extremely predatory so stocking ponds containing tilapia with barramundi of about the same size at a ratio of around 50:1 could prove effective both in preventing overcrowding and in obtaining a valuable bonus. If they could also help to suppress the growth of frogs this would be an additional benefit. Although a marine fish breeding in the sea, they are also found in rivers. They are caught off Hong Kong during the winter which suggest that temperatures in the highlands might suit them.
(v) Other native species
Without expecting very rapid results it would certainly be interesting to get hold of other river and estuarine species from elsewhere in PNG, and look at their culture potential. It should at least be fairly easy to become the world's leading expert on the not-quite-successful aquaculture of some previously untried species (a time-honoured path to an international reputation).
Preliminary experiments on raising fish in ponds which are manured with poultry droppings should be carried out using small ponds, as these will be easier to manage. Manure from ducks, chickens or geese could be tried. Ducks do not actually have to swim on the pond which receives their droppings. Their droppings can be shovelled into the pond regularly, or they can be washed in. Ducks can be confined to a small section at the edge of the pond from which fish are excluded. Or they can be driven in a group from place to place as required. Adult ducks consume about 150–200 g of food per day (dry weight); about 80 percent is later excreted.
Duck rearing could fail for reasons unconnected with the fish such as theft, disease or marketing problems. From the point of view of fishbreeding, chickens, geese or ducks kept off the main pond area would provide manure without reducing the use to which the pond can be put. Ducks which get onto an area of pond where fry can also come may eat many of the fry.
For the future perhaps geese could be considered. Various breeds can be kept in the tropics, especially in moist areas which are not too hot. As grazers they might help to reduce the huge amount of manpower wasted in cutting grass while the droppings produced when they are closed up at night could fertilize the ponds.
Droppings from fed chickens should be as good as those ducks. The advantages would be that duck rearing is virtully untried, whereas there are already quite a number of villagers raising and marketing chickens. Chicken houses can be built over ponds, or manure simply collected from chicken houses and applied as required.
The hatchery is probably not required for carp breeding. Various systems of controlled reproduction are described in Appendix 3, and these are easier to carry out at the ponds than in the hatchery. If, however, Chinese carps are to be introduced and bred the hatchery water supply will have to be improved. At present it has three main problems. First of all it is too cold. It generally comes into the hatchery at 18–22°C, whereas 22–25°C would be better. Secondly the pressure is very variable, and it is not uncommon for the flow to stop completely for a while. Thirdly, although it is generally quite clear it can suddenly become very turbid (S.D. about 10 cm) after heavy rain.
These problems could be partially solved by using as a constant head supply one or two tanks placed behind the hatchery building, up the hill a little way. If these tanks were about 1 m deep and 3 m diameter, each could hold 7 m3 of water. If they were covered on top with transparent plastic sheeting or perspex/plexiglass, then the static water inside would be on average 6–7°C warmer than the present supply. If flowing, the temperature would of course depend on the flow rate and the weather, but two such tanks arranged in series should ensure that a modest flow, (i.e., about 5 l/m) would reach the hatchery day or night warmed by 2–4°C. If the inside of the tank were of a dark colour and the walls insulated (with polystyrene, bubble plastic or straw for example) this effect could be slightly enhanced. The problem of fluctuating pressure would be reduced and a supply held in storage against complete water failures. Although much of the suspended matter in the water is very fine and does not settle even on 1–2 days' storage, there would be some improvement in clarity due to sedimentation. Also of course the worst of the water coming down immediately after rain could simply be diverted away from the tanks in many instances, knowing that one had an adequate stored supply. The insides of the tanks should be made accessible for periodic cleaning to remove silt and algae.
A more complete solution to these problems would be very much more expensive, both to install and to run. It would probably have to involve dosing the water with flocculant and passing it through a sedimentation column, and some form of heating using either heat exchangers or partial recirculation, filtration and sterilization. It would also require more attention and maintenance.
The hatchery is a long way away from the ponds (about 15 minutes walk). Its water supply is limited, unreliable and cold. Induced spawning plus stripping demands skill, reliable systems and dedication, and also some injections during the night. If it is followed by prolonged feeding at high densities it also becomes very expensive. Spawning in ponds or hapas (with or without induction) onto egg-collecting substrates is much simpler. However, the water in almost all ponds is extremely turbid due to clay particles in suspension. This is not helpful to spawning or hatching efficiency. Spawnings carried out in pond 1 recently took advantage of the fact that this pond was without fish and the water was quite clear. It may be that small ponds in which the floor was grassy and no fish were stocked would also provide clear water. This could not be tested. An alternative might be to build a block of concrete tanks in the pond area. They could be fenced for security, but should definitely be without a roof, to take advantage of solar warming. Tanks of about 2 m × 4 m × 1½ m (allowing a water depth of about 1 m) would have a variety of uses.
For spawning such a tank should be cleaned and dried, then filled to about 60 cm depth in the morning and allowed to warm up during the day (perhaps with a clear plastic cover over the top). By late afternoon it should reach temperatures of the mid-twenties, ideal for spawning. Brood fish and spawning substrates could then be introduced, either directly into the tank or into a hapa suspended inside it. A slight water flow to give aeration and a rising water level should help to stimulate breeding activity.
Such tanks should be very useful for short-term holding of various fish such as broodstock after selection or fingerlings before distribution. Before sending fish on long journeys they could be held for a day or two to empty their guts of food, which generally improves survival. Fish from outside the site could be held in quarantine, mullet fry could be acclimatised to fresh water, tilapia larvae could receive hormone diets for sex reversal, diseased fish could be given treatment baths and so on.
The design illustrated in Figs 2 and 3 has the following features.
Tanks receive their water supply via a small channel. This allows a much bigger flow than a piped water supply, and the channel supply at the ponds tends to be warmer than the piped supply. (However, placing the tanks where they can also receive tap water would be an added convenience.) The supply is controlled by boards that fit into slots in the inlet, and can also be screened.
The bottom of the tank has a slight V profile and also slopes down towards the outlet. This should help in collecting very small fish when it is drained. It is important that the tank floor be well made of smoothly finished concrete. Larvae or small fish can also be collected outside the tank by placing a catching net or basket below the end of the outlet pipe inside the drainage channel which should be at least 70 cm wide at these points.
There are two systems for controlling the water level at the tank outlet. The first is by using wooden boards which fit in the outer pair of vertical slots. This would allow any water height to be chosen. A large flow would be possible even with very small fish inside the tank since a big-area screen could fit in the inner pair of vertical slots. As only one layer of boards would be used (unlike a monk where two layers are used and the volume between them is filled with clay) opening and closing the outlet or altering the water height would take only a few seconds. However the boards would leak a little. This would not matter if water was flowing continuously. For static water use (and as a safety measure in case of complete water supply failure) a stand-pipe can be inserted behind the boards. The use of standpipes of different lengths offers a second way of controlling the water height.
In Ayiura pond banks have been made almost vertical in many cases. When erosion occurred attempts were made using wooden posts or stone-filled wire cages to maintain the vertical profile. These were not very successful; wood rotted, erosion occurred behind the posts, wire cages started collapsing inwards. During the present consultancy, filling some of the undercut areas with good quality clay soil was carried out in some areas. The tops of the banks were cut back and the bases built out into the pond, to produce sloping bank profiles. In some cases these were then plastered with pond mud.
These measures seemed helpful. It should become a routine on the site that whenever a pond is drained, the necessary repairs are carried out on the banks. It will soon become clear which methods are effective. In general ponds with larger surface areas and larger fish inside will suffer more erosion due to waves and to the digging activities of the fish. Therefore, the slope of the banks of these ponds will have to be greater.
When ponds are drained, the opportunity should also be taken to remove weeds unless they are required for uncontrolled spawning to take place, to cut back grass along the banks and to dig out sections of the internal drainage channels which may have become full of mud. Other maintenance work particularly for fry ponds is described in Appendix 3.
The OIC, Aiyura is a graduate biologist who has attended a one year multidisciplinary FAO aquaculture course at SEAFDEC in the Philippines. The fisheries technicians have been on study tours to the People's Republic of China, South Korea and Indonesia. None has actually worked on a fully functioning fish farm or hatchery, and they have clearly had difficulty in putting their training into practice.
The practical work carried out with the consultant certainly constituted useful training for the fisheries technician concerned with the ponds and for some of the labourers. Inevitably, however, not all aspects of carp breeding and growing could be dealt with during this period. The OIC was almost entirely occupied with administrative matters during most of the period of the consultancy (and acting as counterpart for the fish feeds specialist during the last three weeks) and therefore was unable to benefit in a direct way. The fisheries technician concerned with the hatchery worked with the consultant to some extent but was away on home leave for much of the itme.
To complement the direct training carried out at Aiyura therefore, the Consultant has prepared a training programme on carp breeding (Appendix 3). It is hoped that this will be helpful to anyone concerned with planning, administering or running a carp distribution station in Papua New Guinea. In addition a section dealing with systems of carp growing is included as Appendix 4. Detailed results and conclusions based on the work carried out while at Aiyura are presented as Appendix 5.
There are at present very few aquaculturists in Papua New Guinea. Therefore those few must be active and fully productive if there is to be any chance of progress. At present there is no doubt that lack of experienced (as opposed to trained) personnel is a serious constraint. Until more practical experience has been gained by local staff, production cannot be expected to rise to very high levels For aquaculturists to spend too much time on administration instead of working in aquaculture is to fail to gain practical benefit from received training and waste a scarce resource.
It would be a pity if the site were to revert to inactivity, and if the technicians and labourers were unable to put into practice what they have learned during the present consultancy. More active involvement at the top and more open discussion between the different levels of staff in the unit would be useful first steps.