WildNothos
THE NOTHOBRANCHIUS SITE
Mafia Island – Notes on the distribution of Nothobranchius and my recent collections
By Béla Nagy
Author’s note: since this article was written, the mainland form of N. korthausae has been described as N. ruudwildekampi by Costa in 2009 and the N. aff. melanospilus from Mafia Island was first identified as a population of N. lucius by Wildekamp, Shidlovskiy & Watters in 2009, until it was described as N. insularis by Costa in 2017.
During January, 2008, I spent some very pleasant, relaxing days with my family on Mafia Island. The island is part of Tanzania and is situated offshore in the Indian Ocean. This is still an almost unspoiled place, with beautiful white sandy beaches. The Mafia Island Marine Park at the southern end of the island is one of the most important protected natural marine habitats on the world. As there are also known freshwater habitats of the genus Nothobranchius on the island, one aim of my visit there was also to carry out a field study of these fishes.
Tanzania is located south of the Equator in East Africa. Mafia is the southernmost of the three bigger offshore islands of the country. It is 48 km long and 17 km wide, and is situated south of Dar es Salaam, near to the Rufiji River delta, about 20 km offshore. The flat-lying island is covered mainly by bush savannah, with some coconut and cashew plantations in the southern part.
Along the eastern coast of Africa the continental shelf averages 8 to 10 km in width, with a notable exception where it broadens significantly to a maximum of 35 km. There, Mafia Island formed by tectonic uplift of the continental shelf. This is an exception along the coast of southern Tanzania where all other islands were formed as a result of extensive coral growth. The terrain on the island is mainly flat; there is a ridge forming a backbone to the northern end but at its highest point it does not exceed 40 m above sea level. Along this ridge there is firm and fertile clay soil where cultivation of annual crops is possible (McClanahan, Sheppard and Obura, 2000; Rubens and Kazimoto, 2003).
The climate experienced by the island is influenced by the northeastern monsoon, the Kaskazi, and the southeastern monsoon, the Kusi. The climate is characterized by periodic rainy seasons. Generally, the longer of the two rainy seasons, the Masika, begins in March and ends in May, and the shorter rainy season, the Vuli, begins in October and ends in mid-December. When the rainy season arrives, the rivers usually overflow their banks and inundate the floodplains. This periodic flooding is essential for survival of the Nothobranchius species, which occur in natural habitats that dry out during the dry seasons.
The early history of Mafia was written by the monsoons, since this natural phenomenon was used to carry the trade ships. Lying on an important ocean trade route of Africa, the island was the site of some very interesting historical episodes. It is mentioned in the ancient Kilwa Chronicles that sons of the first Sultan of Kilwa had settled on the island in the tenth century and this is supported by the presence of ruins on the island. Vasco da Gama had spotted the island on his travels to the north in 1498 and the dominion of the Portuguese lasted for 200 years. On maps of the time the island was referred to as Monfia. Subsequently, the island fell under the reign of the Sultan of Oman in Zanzibar. Around 1820, cannibals of Madagascar invaded the island, and many of the inhabitants were eaten while others were taken as slaves. From 1890 the Germans and later the English ruled over the island, until 1961, when Tanganyika became independent and subsequently merged in 1964 with Zanzibar to form Tanzania.
Nothing of this fascinating history is apparent when the small plane arrives at the primitive airport of the island, after a half an hour flight from Dar es Salaam. Mafia Island is one of the least developed parts of a poor country. There are no tarmac roads, only two main dirt roads and some minor sandy roads or tracks. Most of the forty thousand inhabitants work in agriculture or fishing.
The southern part of the island
The first news about an undescribed species of Nothobranchius from the island came as a result of a night-time frog-collecting exercise carried out by the German Dr. Herkner sometime in January, 1972. Besides the amphibians he sought, he also collected eight small cyprinodonts. The collector had visited the island, mainly for diving purposes, in the company of Edith Korthaus who, at that time was editor in chief of the German magazine Das Aquarium. The collection site was a large swampy area of some hundreds of square meters in extent that apparently had not dried up completely during the dry season. A pathway crossed this swamp with brown muddy pools on both sides. Unfortunately, only one male specimen survived the trip back to Europe (Korthaus, 1972).
One year later, Korthaus again visited the island, this time with Gerd Eggers and Christel Kasselmann. They again found the species at the same locality, in small depressions alongside the beaten track. Based on specimens from this collection the original description was made by Hermann Meinken (1973), and the species was named in honor of one of the discoverers as Nothobranchius korthausae. The species was also found at some additional locations in larger bodies of water. Interestingly, the water of all the habitats was acidic, with an average pH of 5.8 and a maximum of 6.4. The species was not found at localities where the pH of the water was higher than 6.4 and, according to Korthaus (1973) it was not even possible to use such higher pH water for water changes when the specimens were kept in plastic bags.
The type locality was situated just about one kilometer to the north of the Mafia Island Lodge, near Utende (Eggers, pers. comm.). Eggers and Kasselmann, together with Lothar Seegers, again collected the species from the type locality and in the surrounding area late in 1980. Further, Eggers and Kasselmann collected it yet again, one year later (Kasselmann and Eggers, 1982).
When breeding the species at home, Eggers observed that one specimen developed darker coloration than the others. He organized another collecting trip, in company with Kasselmann and Walter Scholl, in an effort to learn if another color form existed on the island. On that trip they again found the original yellow color form at the type locality, and indeed, at an additional location a red variety was found. The locality was a large, shallow depression beside the main road from Utende to Kilindoni, about 5 km west from the Mafia Island Lodge, still in the southern part of the island. Seven yellow and five red males were found. Unfortunately, despite their best efforts no females could be collected from that location. The males were taken home and bred with females derived from the earlier pure yellow aquarium strain. So, at that stage, the strain in the hobby represented a cross between the red and yellow forms from two different collections (Eggers, 1982; Eggers, pers. comm.).
During our stay on the island in January, 2008, the dry season was well advanced in the southern part of the island. Walking from the Mafia Island Lodge, I tried to locate the area of the type locality. Unfortunately, all the potential habitats, depressions at culverts, and any other depressions that could contain Nothobranchius during the rainy season, were totally dry. The bushy grassland was relatively easy to explore because numerous pathways crossed the area. Those paths often led to houses and I asked the inhabitants where I might find some water in the area. I always got a negative answer regarding the presence of rest water there. However, they assured me that the central and northern part of the island was still wet and that I would certainly find some water there.
While walking in the area, a loud clanging noise caught my attention. It turned out that a large and noisy colony of Pteropus bats were resident in the tall trees near a mangrove swamp. This fruit-eating species could also be seen in great numbers at dusk, when flying out to search for food. According to the scientists, further research may reveal that the group on Mafia Island is a distinct species.
The central and northern part of the island
The early reports described localities of Nothobranchius korthausae mainly in the southern part of the island. During 2002 a collecting trip was carried out by Brian Watters, Barry Cooper and Otto Schmidt, and they were successful in finding Nothobranchius fishes, not only in the southern part of the island but also in the central and northern parts. In the eastern and western central part they found both the yellow and red varieties of N. korthausae, in some cases living sympatrically in the same biotope. The northern part of the island presented some additional interesting discoveries. At two locations they found an undescribed species that is presently referred to as N. aff. melanospilus. At one of the two sites, this species occurred together with both the red and yellow forms of N. korthausae, as well as an interesting intermediate, bicolored, form of that species (Watters, pers. comm.).
My second excursion was to try to collect from the central and northern parts of the island. For various reasons, it is not easy to rent a vehicle on Mafia Island. There is an absence of tarmac roads and only a few cars on the island, all of them are generally booked for the transfer activities between the airport and the few hotels. There are also no vehicle maintenance facilities available and the well-used old cars there often break down. Fuel is in short supply and very expensive. After some organization I succeeded in hiring a car and a driver for one day. With the lack of a public transportation system on the island, we often had to give lifts to local people. Starting from Utende in the southeastern part of the island, we drove first in the direction of Kilindoni. This is the dirt main road that connects the main town on the west coast with the hotels in the area of the marine natural reserve on the east coast. According to the driver and also one of our newly-acquired “fellow-travelers”, there are usually great areas along the road that are under water during the rainy season. However, at the time of my visit this area along the main road was completely dry.
We soon arrived in Kilindoni, the main administrative center of the island. After some very complex barter arrangements, that I did not fully understand, we were able to obtain some fuel in two plastic containers. In exchange we gave other items stored in boxes in the back of the car. The content of one container was immediately poured into the fuel tank and the other was placed in the back of the car, right beside my collecting equipment. I tried to ensure that there was always enough distance between them, in order to avoid contamination of the collecting equipment, but it was not easy with all the shaking on the way.
After obtaining the fuel, we drove north on the main road in the direction of Ndagoni. Before reaching that village we stopped at a place where the bush alongside the road was rather thick but through it we could see the shimmering of water.
About 10 meters from the road there was an open pool, with violet-colored water lilies, in the middle of a large swampy area. It was difficult to approach the open part of the pool because the edge of the swamp was densely vegetated with reeds. However, my efforts were soon rewarded with a beautiful male specimen of the yellow form of Nothobranchius korthausae. Soon after, I collected the first female. They were relatively large specimens with good color, especially the light blue margin on the unpaired fins of the males. The fish were bagged individually and we left the location. (Location Mafia Island TZN 08-2)
After driving some hundred meters further on the same road we arrived at a small “lake”, situated directly beside the road. In that large pool there were, again, violet-colored water lilies. This pool measured about 50 meters in length and there was a culvert under the road connecting it with a small muddy pool on the other side. In that small muddy pool, which was very shallow and at the point of drying up, I found a lot of cichlids of about 6-8 cm in length, most probably of the species Sarotherodon urolepis. The water was, at that stage, so shallow that the backs of the fishes were above the water surface. When I stepped to the edge of the habitat they took fright and jumped out of the shallow water. I then turned my attention to the larger body of water, which was clear, with cichlids swimming in large groups among the vegetation. Although my previous experience had indicated that the occurrence of Nothobranchius is unlikely if the cichlids are abundant, I nevertheless wielded my net and soon found a yellow male of N. korthausae. Attempts to collect further specimens in different parts of the habitat were unsuccessful. While I was happy to see that there was a cyprinodont living in the habitat in spite of the large number of bigger fish, clearly the presence of the cichlids had had a negative effect on the size of the Nothobranchius population. (Location Mafia Island TZN 08-3)
When we left this location we turned off the main road onto some narrow, sandy and winding minor roads. Finally we arrived at a place the driver had been aiming for. In the middle of dense bush there was a depression in the road that was flooded with water to a depth of about 5-10 cm. There was no aquatic vegetation, only some decaying leaves on the bottom. On one side of the road the water extended under the bushes. When I stepped into the small pool in the road, I saw some small fish trying to hide under the leaves on the bottom. At the side of the pool, among the roots of trees, small fish were abundant. They were half-grown, but already clearly sexable, specimens of the red form of N. korthausae. Reports of earlier collections indicated that the red form had always been found with the yellow form in the same biotope. My small plastic container was quickly filled with fish, but I continued to collect more specimens in an effort to determine if both color forms were present in the habitat. Finally, I came to the conclusion that only the red form was present. After selecting six trios and bagging them individually, I released the remainder into the habitat. This might be the first confirmed habitat where the red color form only was found. The males are a bright solid red with hardly any markings. Further breeding experiments will show if any yellow specimens will turn up among the offspring. (Location Mafia Island TZN 08-4)
Returning to the main road, we passed through the village of Kirongwe, heading northward. Soon we arrived at the location where Watters, Cooper and Schmidt had collected N. aff. melanospilus in 2002 (Watters, pers. comm.). Beside the road, an area of dark clay showed where there had been a pool which had recently dried up. A short way off, behind some trees, there was still some very shallow water that would, no doubt, have been part of the same seasonal pool system during the rainy season. The pool was most likely the last remaining drinking place for the cattle in that area as evidenced by the presence of a huge amount of cow dung. Although the water was only about 5 cm deep, the mud substrate was very thick and when I stepped into the pool I immediately became stuck in it. In the very shallow and totally turbid water it was not possible to move the net without also scooping up a lot of mud. In spite of these conditions I caught a large number of N. aff. melanospilus with every sweep of the net. In the plastic container we could not see the fish because the water looked like a mud-soup. The fish was half-grown, without bright coloration and I tried to separate the males and the females on the basis of fin shape and traces of some coloration. The pale coloration was probably influenced by environmental factors as they developed beautiful colors shortly after being placed in aquaria. The fish were in good physical condition, which was more than surprising, given the small amount of water, the high population density, and a probable lack of sufficient food. Furthermore, huge tadpoles were very abundant in the pool. These predatory tadpoles were at least double the size of the fish and they had large, broad mouths. In a few minutes I collected more fish than I needed, so I selected and bagged ten pairs, returning the remainder to the habitat. It would appear that the short rainy season had brought less water than usual. The shallowness of the habitat at that time indicated that the pool would have dried out very soon after my visit and, in such a short time; the fish could not be expected to reach full size. However, as the fish were already sexed out, they must have been spawning, thereby ensuring the survival of the population. This was confirmed when the fish started to spawn shortly after arrival in aquaria after traveling a few days in the bags. (Locality Mafia Island TZN 08-5)
We then headed back southwards, passing through Kirongwe, to a further location where Watters, Cooper and Schmidt had found another population of N. aff. melanospilus, but in that case co-existing with the yellow, red and intermediate forms of N. korthausae. Although the habitat still contained some water, there were no fish to be found there. It was possible to see the substrate of the pool through the slightly turbid water. Beside the pool there was a kingfisher, with a full belly, perched on a tree!
Continuing southward on the main road, we again arrived at Ndagoni. Immediately south of the village we turned off onto a small road that represented a short-cut back to the hotel in Utende. This also provided us with the opportunity to search for further habitats. The narrow sandy track passed periodically over watercourses with culverts under the road. On both sides of the culverts there were depressions, some still containing water. We always stopped at such points and looked for fish. I almost always found cichlids and other non-annual fish, mainly in such habitats where the water was clear. At one location I succeeded in finding the yellow form of N. korthausae again. In this habitat the water was completely turbid. When I put my net under the roots in the water I also caught some large catfishes. I tried to get rid of them immediately before they could damage my collecting equipment while jumping around with their thick, heavy bodies in the net. After bagging the specimens of Nothobranchius we returned to the hotel. A few hours later I observed that the fish from this last locality were in poor condition, all of them with clamped fins. They probably had some kind of bacterial infection and looked ill until they were placed in aquaria after the journey. There, they recovered completely within 24 hours. (Location Mafia Island TZN 08-6)
Observations and conclusions
Distribution
Mafia Island is of recent origin and was formed during the Pleistocene epoch when carbonate deposits were laid down while the island was raised above sea level through faulting of the continental shelf (IPP Media, 2008). The Pleistocene was characterized by repeated glacial and interglacial cycles that lasted for some ten thousand years each. During the last glacial period of the ice age, about 20,000 years ago, the sea level was about 120 meters lower than it is today. This exposed the continental shelves and formed land-bridges between the landmasses, making the migration of animals possible. N. korthausae is also found on the mainland opposite Mafia Island, mainly around the Mbezi River (Wildekamp, 2004). When the sea level was much lower the species probably inhabited a larger area. The rise of sea level isolated the populations of the species on Mafia Island. According to the available information of the earlier collections, complemented with the recent collections, it would seem that N. korthausae occurs throughout the island, in habitats formed on the clay soil along the central ridge.
The second species that is known from the island, N. aff. melanospilus was only found in the northern part of the island at two locations within a relatively small area. This species is also known from the mainland in two fairly restricted areas that are quite widely separated, the drainage system of the Kilombero River and in the vicinity of the Mbezi River.
Color forms
N. korthausae is known to have two color morphs, yellow and red. On Mafia Island the yellow form is widespread and occurs throughout the distribution range, in some cases existing sympatrically with the red form and, in at least one locality, with an intermediate form. There are some differences in the markings of the different populations of the yellow forms, even though the distance between the locations is relatively small. The red form occurs most typically in the central and northern parts of the island. It has, however, also been reported from the southern part of the island (Eggers, 1982; Eggers, pers. comm.) but this could not be supported by personal observation.
Water conditions of the habitats
In the original description of N. korthausae by Meinken (1973) an observation was made regarding the pH values of the habitats. It was stated that the species lived only in acidic water, and it is not possible to maintain them in alkaline water (Korthaus, 1973). This theory was quoted further in several other publications, especially in Germany.
The entire island consists of coral rag overlain by sandy loam soil [Editors note: Coral rag is a rubbly limestone composed largely of ancient coral reef material]. The soils are mainly alkaline with an average pH of 7.5 (IPP Media, 2008). Because of the buffering capacity of the substrate the water of the habitats is mainly alkaline. My personal observation on the island was that most of the habitats had alkaline water, and Watters, Cooper and Schmidt also noted the same (Watters, pers. comm.). If the water of the natural habitat was acidic, it was essentially because of the presence of a large quantity of decaying vegetation. During the transport of the fishes on the island and also for their maintenance in aquaria, I have used water with mildly alkaline conditions. I did not experience any losses when using alkaline water on the island or at home.
Seasonality of the habitats
In the original description it was mentioned that the habitats usually do not dry out totally during the dry season. However, in January, 2008 I found the entire region near the type locality to be totally dry. My driver took me to some localities which normally contained water, but those were also dry. Where we collected some fish in the central and northern part of the island, the water level was almost always very shallow and at the point of drying up completely. As we were there at the beginning of the dry season, it is reasonable to assume that by the end of that dry season nearly all the habitats would have dried completely.
Localities
N. korthausae Mafia Island TZN 08-2. A pool of 10 x 20 m in size, consituting the open central part of a extensive swampy area, heavily vegetated with water-lilies and reeds. On the northwest side of the main road from Kilindoni to Kirongwe, 1.4 km west of Ndagoni and 9 km northeast of the airport at Kilindoni. The water was clear, light brown and had a maximum depth of 1 m. The water temperature was 29.3 °C, the pH was 6.14, and it contained 51 ppm total dissolved solids. Date of collection: January 6, 2008.
N. korthausae Mafia Island TZN 08-3. A large pool of about 50 m in length, in part heavily vegetated with water lilies and areas of reeds. On the south side of the main road from Kilindoni to Kirongwe, 0.6 km west of Ndogoni and 9.8 km northeast of the airport at Kilindoni. The water was clear, light brown and had a maximum depth of 1 m. The water temperature was 33.1 °C, the pH was 7.21, and it contained 52 ppm total dissolved solids. Date of collection: January 6, 2008.
N. korthausae Mafia Island TZN 08-4. A pool formed in a depression across a small road, about 5 m in diameter, extending in part under bushes on one side of the road. 3.2 km southwest of the junction with the main road in Kirongwe. The pool lacked aquatic vegetation, and had only some decaying leaves on the bottom. The water was clear and light brown, about 5-10 cm deep on the road, and about 20 cm deep at the deepest part under the bush. The water temperature was 28.7 °C, the pH was 5.79, and it contained 56 ppm total dissolved solids. Date of collection: January 6, 2008.
N. aff. melanospilus Mafia Island TZN 08-5. A rest pool 6 x 4 m in area, situated partly in the shade of a large tree. On the south side of the main road from Kirongwe to Ras Mkumbi, about 1 km northeast of Kirongwe. The water was muddy and about 5 cm deep, with a very thick mud substrate. The water temperature was 33.3 °C, the pH was 7.59, and it contained 411 ppm total dissolved solids. Date of collection: January 6, 2008.
N. korthausae Mafia Island TZN 08-6. Pools on both sides of the road, each about 10 m long, connected through a culvert under the road. About 3.3 km north of the main Utende-Kilondoni road from a junction located about 4.3 km west of the Mafia Island Lodge in Utende. The pool on one side was more open and vegetated with water-lilies, the pool on the other side was among bush and without aquatic vegetation. The water was brown and turbid. The water temperature was 31.6 °C, the pH was 7.50, and it contained 93 ppm total dissolved solids. Date of collection: January 6, 2008.
Note that location code designation TZN 08–1 was assigned to a population of N. neumanni on mainland Tanzania that was collected prior to travelling to Mafia Island. For that reason, the details of that site have not been included in the above listing.
Acknowledgements
Firstly, I would like to thank my family for allowing me to do the field study and collecting, while we were on vacation. I want to express my special thanks to Susanne Schories and Manfred Schartl (Würzburg, Germany) for the official documents and Brian Watters (Nanaimo, Canada) for valuable information about localities on the island and for reviewing this manuscript.
References
Eggers, G. 1982. Der rote korthausae von Mafia. Das Aquarium 16 (162): 632-633.
IPP Media. 2008. Mafia Island Marine Park to protect marine resources: http://www.ippmedia.com/ipp/guardian/2008/08/30/121575.html
Korthaus, E. 1972. Killifisch, noch ungetauft. Das Aquarium 6 (40): 10–12.
Korthaus, E. 1973. Bemerkungen zum Biotop von Nothobranchius korthausae spec.nov. In: Meinken, H. 1973. Nothobranchius korthausae spec. nov., eine hübsche Cyprinodonten-Neuheit von der Insel Mafia (OstAfrika). Das Aquarium 7 (51): 351-355.
Kasselmann, C. and G. Eggers. 1982. Sansibar und Mafia, sieben Jahre danach (II). Das Aquarium 16 (160): 518-522.
McClanahan, T.R., C.R.C. Sheppard and D.O. Obura (eds). 2000. Coral Reefs of the Indian Ocean: Their Ecology and Conservation. Oxford University Press, New York: 552 pp.
Meinken, H. 1973. Nothobranchius korthausae spec. nov., eine hübsche Cyprinodontiden von der Insel Mafia (Ost Afrika). Das Aquarium 7 (51): 351–355.
Rubens, J. and S. Kazimoto. 2003. Mafia Island a Demonstration Case. Application of the WCPA-Marine/WWF Guidebook on Evaluating Effective Management in MPAs. September 2003.
Wildekamp, R.H. 2004. A World of Killies. Atlas of the oviparous cyprinodontiform fishes of the world. Volume IV. American Killifish Association, Mishawaka, Indiana: 368pp.
Additional references since publication
Costa, W. J. E. M. 2009. Species delimitation among populations of the eastern Tanzanian seasonal killifish Nothobranchius korthausae (Cyprinodontiformes: Nothobranchiidae). Ichthyological Exploration of Freshwaters, 20 (2): 111-126.
Wildekamp, R. H., K. M. Shidlovskiy & B. R. Watters. 2009. Systematics of the Nothobranchius melanospilus species group (Cyprinodontiformes: Nothobranchiidae) with description of two new species from Tanzania and Mozambique. Ichthyological Exploration of Freshwaters, 20 (3): 237-254.
Costa, W. J. E. M. 2017. Redescription of Nothobranchius lucius and description of a new species from Mafia Island, eastern Tanzania (Cyprinodontiformes, Aplocheilidae). Zoosystematics and Evolution, 93 (1): 35–44.