STEM CRACKS AND DECAY IN NEWTONIA BUCHANANII TREES IN THE MAZUMBAI FOREST RESERVE, WEST USAMBARA MOUNTAINS, TANZANIAF.A. MremaTanzania Forestry Research Institute (TAFORI), Silviculture Research Centre,P.O. Box 95, Lushoto, TanzaniaM. NummelinFORST ProjectP.O. Box 292, Morogoro, Tanzaniaforst@twiga.com&Division of Environmental Biology, Department of Ecology and SystematicsFIN-00014 University of Helsinki, FinlandMatti.Nummelin@formin.FIABSTRACTStem cracks, decay and bark pattern in Newtonia buchananii trees were investigated and compared in a submontane rain forest in the Mazumbai Forest Reserve and adjacent human disturbed forest. One third of the trees growing in the reserve (32 %) and more than half of those in the disturbed forest (60 %) had stem cracks and/or decay-mainly as butt rot, often accompanied by brackets of wood-rotting fungi. One fifth (20 %) of the trees growing in the reserve and three quarters (76 %) in the disturbed forest showed rough and rugose bark. As N. buchananii is a main climax species in the area, the outcome of the death of the trees will be the formation of big gaps. Gap formation in the reserve seems to be accelerating and going beyond the 'normal' levels of a natural rain forest ecosystem, whereas in the disturbed areas it seems that N. buchananii will be extinct in the future.INTRODUCTIONNewtonia buchananii (Baker) Gilb & Bout (Leguminosae, Mimosoidae) is a dominant high canopy tree species, widely scattered in submontane rain forest in Tanzania. It is commonly distributed within the Eastern Arc Mountains, especially in the Usambaras, Nguru and Uluguru mountain forests. N. buchananii timber is valuable and suitable for general joinery in building and construction work (Bryce, 1967).Recently, stem cracks and decayed wood was detected in living N. buchananii trees growing within Mazumbai Forest Reserve and in the human disturbed forest surrounding the Reserve. While the extent of the damage appears to be increasing in both forest types, the causes of the cracks and decayed wood in N. buchananii are not yet known. There is very little information available about damage to N. buchananii. Struhsaker et al. (1989)investigated tree mortality in the Kibale Forest, Uganda, and noted a more severe dieback due to an unknown cause(s) in N. buchananii than in any other tree species in their study.It is striking that besides N. buchananii, a climax tree in the Forest Reserves, another abundant climax tree, Ocotea usambarensis (East African camphor tree) is also declining seriously (Nsolomo & Venn, 1994). Heart rot in stems of O. usambarensis has been reported and the extent of the degradation, by the pathogens, was enormous in every locality where this species grows (Dick, 1969; Renvall & Niemelä, 1993; Nsolomo & Venn, 1994). In addition, Nsolomo & Venn (1996), while studying decay of the O. usambarensis, reported stem cracks as a symptom and sign of a tree suffering from internal stem decay. However, such cracks were only common in trees with extensive butt rot in which the sapwood was also affected. Renvall & Niemela (1993) reported 10 polypore species (Basidiomycetes) all causing wood decay in O. usambarensis growing in the West Usambara Mountains. Four of those species were noted to attach to living trees. Nevertheless, the mode of infection of wood inhabiting fungi into naturally growing tree species is not clearly documented although several studies have demonstrated stem cracks as entry points for pathogen infection into living trees (Bostock & Stermer, 1989).Stem injuries, enclosed branch stubs and basal sprouts have been reported as primary causes of stem cracks in living trees (Butin & Shigo, 1981), while drought stress is a triggering mechanism of the cracks (Day, 1954). Butin (1995) explained that drought could induce stem cracking in old cambial injuries on living stems, as they are weak points prone to crack during extreme moisture tension in a tree. However, in Scandinavia, Dietrichson et al. (1985) and Persson (1994a) reported drought as a triggering factor for the cracks in fast growing Picea abies (Norway spruce) with large annual ring width variations, but without any noticeable cambial injuries.Most of the published studies (see above) have been done in temperate regions. Stem cracks in trees growing in tropical submontane forests are poorly studied. In general such investigations in tropical forests seem to be very scarce.This study was conducted at the Mazumbai Forest Reserve and in the adjacent human disturbed forest. The distribution of wood defects (stem cracks and decayed wood) in N. buchananii trees was investigated. The main objective was to quantify the extent of stem cracking in N. buchananii according to tree sizes, with the intention of studying the possible causes of stem cracks, and ultimately the occurrence of pathogenic infection in N. buchananii trees.MATERIALS AND METHODSStudy siteThe study was carried out at Mazumbai Forest Reserve (4°50'S and 38°30'E) and in surrounding human disturbed forests. This Reserve is one of the few remaining intact submontane forests located in the West Usambara Mountains (figures 1A, B).To the South and West of the study site there is Baga Forest Reserve and on the North is a tea estate, which is owned by the villagers. On the eastern side is the Sagara Forest, which is managed by local people. Most of the Sagara Forest is highly disturbed and a part of it has been cleared for cultivation. Common climax canopy tree species in the Mazumbai Reserve are Newtonia buchananii, Ocotea usambarensis, Sorindeia madagascariensis and Syzygium guineense. The Reserve covers an area of about 320 ha and is managed by Sokoine University of Agriculture for training and research purposes.Figure 1. A. Map of West Usambara Mountains in NE Tanzania. B. The Mazumbai Forest Reserve showing the transect lines in the Reserve and in human disturbed forest.Climatic conditionsMean annual rainfall recorded for nine years (1988-1996) is 1,200 mm. The driest year, with rainfall of 874 mm was 1992, while the wettest year was 1989 with 1,954 mm of rainfall. There are two dry seasons with monthly rainfall less than 100 mm (figure 2). These are January-February and June-September.The yearly numbers of the drought and dry months show an increasing trend (figure3 A, B). Although the analysis period was short, the pre-humidity index calculated according to Walsh (1996) is 3.5, suggesting a substantial climatic change, having now tropical wet-dry conditions with decreasing relative humidity (Walsh, 1996). The index also indicates that the drought conditions were severe enough to create water deficit in trees within the tropical habitats (Walsh, 1996).Method of assessmentA total number of 104 plots of 0.1 ha were demarcated, 52 within each forest type (Mazumbai Forest Reserve and adjacent human disturbed forest). A total of 261 N. buchonanii trees were examined to investigate the occurrence and distribution of stem cracks and decay in the two forest types.Figure 2. Mean monthly rainfall (mm) at the Mazumbai Forest Reserve (1988-1996).Morphological features, extent of the stem cracks along the boles and the degree of decay were examined in each tree in the plots. The condition of the tree was recorded as healthy (for healthy-looking trees without any external damage like wounds, cracks and/or decay: with reference to Shigo, 1983a); and as cracked and/or decayed (for those trees with open or occluded cracks, decay and/or with bracket fungi). Where cracks were not occluded, it was easy to associate their presence with internal decay or cracking due to fluttering of the buttresses. Decayed wood in the examined tree was confirmed by the presence of brackets of wood rotting fungi or by scratching the rotten wood with a knife and feeling it between the fingers. Bark patterns (categorised as smooth, rugose and rough) were recorded as described by Junikka (1994). Thorough inspection was conducted on every N. buchananii tree to investigate the incidence of damage and possible causes. The occurrence of brackets of wood rotting fungi on the trees were recorded and associated with diameter classes for comparison between the reserve and the disturbed forest.Figure 3. A. Monthly rainfall and mean temperature at Mazumbai Forest Reserve (1988-96). B. Number of months with drought condition (rainfall ⟨ 50mm/month) and dry condition (rainfall ⟨ 100mm/month) showing the trend between 1988 and 1896).Data analysisData were statistically analysed using Systat (version 5.0) (Wilkinson, 1992). Comparisons of the distribution of the defective trees were made using a Chi-square test. Comparisons of bark patterns in both forest types were done in the same manner.RESULTSStem cracks and decayStem cracks and decay were observed along the boles and on some large branches of N. buchcmanii. The defects of trees growing in human disturbed forest are shown in figure 5 (A-B), while those in the reserve are as shown in figure 5 (C-D). The cracks in the human disturbed areas were mostly wide open, while those in the reserve were often occluded. One third (32 %) of the 140 trees inspected in the reserve showed stem cracks and/or decayedwood, while 60 % of the 121 trees investigated in human disturbed forest had same types of defects (table 1).Table 1. Tree conditions of N. buchananii trees as observed in the Mazumbai Forest Reserve (140 trees) and in the adjacent human disturbed forest (121 trees).Forest typesTree conditionBracketfungiBark patternHealthyCrackedPresentAbsentSmooth Rough/ rugoseMazumbai Reserve95453210811129Disturbed forest487327942893The difference in the occurrence of defects between the habitats is statistically significant (χ⟨sup⟩2⟨/sup⟩=20.2, p⟨0.001). There were no healthy trees growing in the human disturbed forest with diameters ranging from 130 to 160 cm (figure 4). The defects in disturbed forest occurred even in young trees of diameter class 40-60 cm. In the Reserve damage occurred only in mature trees (⟩70 cm). The study also revealed that healthy and defected trees are not equally distributed in size classes. Large diameter trees showed a higher frequency of stem cracks and decay (χ⟨sup⟩2⟨/sup⟩ = 39.6 and 46.9 for the reserve and disturbed forests respectively, p⟨0.001 for both).Figure 4. Occurrence of stem cracks and decay in N. buchananii trees associated with diameter classes.Bark patternsMost of the trees (77 %) growing in human disturbed forest had rough and rugose bark (with weak scaling, wrinkles, depressions and shallow fissures) (figure 5A, B) while those in the reserve had smooth bark (79 %). However, some trees in the reserve had rugose bark patterns that looked smooth from a distance of three to five meters (figure 5C, D). Bark patterns in human disturbed area were more rough and rugose and only a few showed smooth bark, while those within Mazumbai Reserve were generally smooth and only few had rugose bark (table 1). The difference is statistically significant (χ⟨sup⟩2⟨/sup⟩=82.19, p⟨0.001).Bracket fungiBrackets of wood rotting fungi were common on N. buchananii trees. Most of these polypores were noted on the bark of living trees between buttresses. Only a few were observed on the branches. Larger diameter trees had most of observed brackets. Also, one small tree (diameter about 50 cm) growing in disturbed forest had a bracket fungus developing at stump height. This indicates that even young and otherwise healthy-looking trees could be decayed.DISCUSSIONStem cracks and bark patternStem cracks in tree species growing in temperate conditions have been studied extensively (e.g. Caspari & Sachsse, 1990; Day, 1954; Dietrichson et al., 1985; Kubler, 1987; Neely & Himelick, 1987; Persson, 1994a, b). Some of these studies showed that previously wounded trees were often cracked later on (Kubler, 1987). However, such studies are virtually nonexistent for rain forest trees.In this study, two types of cracks were observed: the first type was those cracks which occurred due to the presence of internal wood defects such as decay or old cambial injuries (Shigo, 1983a). Caspari & Sachsse (1990) suggest a physiological explanation for stem cracks occurrence: during prolonged dry periods, excess transpiration suction stress could cause cell rupture in low strength wood and that such damage might lead to stem crack development in trees. A trend towards more extreme and harsh climatic conditions, such as those observed at Mazumbai over the past few years could be the cause of the observed stem cracks in many of the already internally defected N. buchananii. Also, O. usambarensis growing in the proximity of N. buchananii were reported to have developed wood decay in living trees some years back (Dick, 1969; Nsolomo & Venn, 1996). Stem cracks were scantily reported and only as a symptom and sign of an extensive butt rot that involves the sapwood (Nsolomo & Venn, 1996). The second type of stem crack was noted to probably occur during buttress formation. This suggests that dilatation processes, which are caused by the tangential strain as a result of stem diameter increase (Trockenbrodt, 1990; Junikka, 1994), or enlargement of the buttresses, could cause stem cracks even if the tree was not internally defective. No attempt has been made so far to investigate the timing of the appearance of such cracks. Persson (1994a, b) reported that most stem cracks in P. abies developed during periods of late-wood formation, and the cracking was triggered by hot dry weather conditions. Stem cracks in Cupressus lusitanica, reported by Mrema (1997) are an example of drought problems in Tanzania. He suggested that mechanical injuries were the primary cause of the observed cracks, which usually appeared after prolonged drought.Figure 5. Stem cracks and decay in a N. buchananii growing Mazumbai Forest Reserve and in human disturbed forest. A. Trees in a human disturbed forest showing decay all along the bole. B. A tree in the human disturbed forest with rough and rugose bark. Note the open crack. The tree has previously been wounded at the base. C. A tree in the Mazumbai Forest Reserve. Note the occluded crack. D. A tree in the Mazumbai Reserve showing smooth bark despite the internal decay. Note the decay was contained at the base as butt rot.The trend in the climatic conditions for the past nine years (figure 3B) shows that drought periods over the last five years (1992-1996) were even more pronounced than those reported 20 years earlier in Tanzanian highlands (Lundgren, 1978). Long drought periods have an extreme ecological impact on rain forests (Walsh, 1996). Mean monthly rainfalls during the nine years at Mazumbai showed that there were about six months of dry weather (June-November, figure 2). According to Kramer (1983), water deficit develops more in shallow rooted tree species like N. buchananii and can cause cracking in trees with internal defects. It is likely that O. usambarensis trees do not develop stem cracks so often because their roots grow deep into the soil and are able to find water during drought conditions.Most of the stem cracks on trees in the Reserve were occluded (figure 5C), while those within disturbed areas were open (figure 5B). This indicates that trees in the closed forest probably have a more humid and cooler environment, which is better for wound healing. Another effect of a different microclimate could be a different bark pattern. In the Reserve, the bark of most of the trees was smooth. The rough bark, observed on trees growing in disturbed forest, suggests that the trees were more exposed to drier conditions and probably to higher temperatures. This could hinder the ability of cracks to occlude. High temperatures can directly damage trees by killing the bark and are referred to as sun scorches (Butin, 1995). The observed rough bark pattern in N. buchananii trees might illustrate symptoms of water deficit development in the trees.Causes of injury and decay on treesGenerally, there are three well-known forms of bark damage: mechanical wounding, injury from extreme weather, and biotically induced necrosis (Butin, 1995). In this study, the injuries of N. buchananii trees growing in the Mazumbai Reserve were mainly caused by tree fall or snaps, branch falls and in some cases animals like bush babies (Galago spp) chewing the bark to reach sap. Human damage was found on N. buchananii trees growing in the human disturbed forest. Apart from logging damage, the injuries may have occurred as the bark of this tree is used by local communities as a medicine for livestock. Apart from mechanical wounding, fire was noted to be a menace to trees growing in disturbed forest. During farming seasons, uncontrolled fires often cross the forest boundary, damaging the bark and creating numerous other injuries on the stems. As soon as the tree is wounded, the living cells of the sapwood respond to the injuries through chemical changes thus causing wood discoloration. Such discoloration in wood is an incipient of wood decay in living trees (Shortle, 1984). It is well documented that wound injuries are the primary causes of wood decay in living trees (Bostock & Stermer, 1989; Nsolomo & Venn, 1996; Pearce, 1996; Shigo, 1979, 1982; Tate, 1986; Tippett & Shigo, 1981). However, Dick (1969) did not find many decayed young trees of O. usambarensis, even when there was bole damage or broken branches. Instead, decay was common in most of the mature O. usambarensis.The observation of the fungi and decay between the buttresses indicated that there was a pathogenic infection mainly causing butt rot. The bracket fungi were also observed to attach on the bark of small diameter trees indicating that the infection on living trees, apart from being through the stem injuries, could be for example via the root as it was observed in one of the studied trees in Mazumbai Reserve. The observed bracket fungi are presumably Phellinus species (Niemelä & Nsolomo, pers. comm.). Phellinus senex is reported by Renvall & Niemelä (1993) to grow on root systems of huge living trees of O. usambarensis while Loveporus inflexibilis is one of the main decay-causing fungi of O. usambarensis. Taking into consideration that O. usambarensis has an extensive rooting system and that these fungi are root pathogens of large old trees causing root rot and butt rot, it could be possible that the pathogen infection in N. buchananii originated from close roots of the two species as they all grow in the same habitat. Nsolomo & Venn (1996) hinted that symptoms of the presence of the root rot in trees are butt rot and stem cracks, as we observed in N. buchananii, hence our hypothesis that the pathogen infection in N. buchananii occurred also through the roots was substantiated.As N. buchananii is the major climax tree species in the area, the consequence of decay and death of a great portion of the trees will be the formation of big gaps. The gap formation process might be more than a regular process, as at the same time another climax species O. usambarensis is declining in that area. Sudden big changes especially in the Reserve will probably favour growth of many under-storey canopy species. Gap formation in the Reserve seems to be accelerating and maybe going beyond the normal rate of occurrence of healthy rain forest ecosystems. Whereas in the disturbed areas, it seems like N. buchananii will be extinct as a timber tree in near future.This study indicates the need for further investigations of disturbance ecology in the natural forests in Tanzania. The dynamics of host-pathogen environment interactions of the major rain forest trees needs to be clarified. The extensive rot in N. buchananii is a problem, which could be man-made. The high incidence of rot in the disturbed area could be a source for spread into the Reserve. The parasites of N. buchananii should also be identified. Finally, studies are needed on the spread of rot within trees and between trees when old trees are falling and gaps are forming, when spread to younger trees is supposed to occur. Population biology studies, both in the disturbed area and in the Reserve could answer the questions about how the rot spreads and can increase the knowledge about the ecological effects of wood rotting fungi in tropical forests.ACKNOWLEDGEMENTSWe gratefully acknowledge the support of the Finnish Forestry Research Support in Tanzania (FORST) Project. Initial studies of FM about stem cracks in living trees were funded by the International Foundation for Science (IFS), the knowledge that led to formulation of this study. Comments by M. Ruokonen improved the earlier versions of this manuscript. We are sincerely grateful to Pia Barklund for valuable comments on the manuscript. Many thanks to Sokoine University of Agriculture, Faculty of Forestry for allowing us to conduct the study at Mazumbai Forest Reserve and M. Mrecha, for providing us with the climatic data.REFERENCESBostock, R.M. & B.A. Stermer (1989). Perspectives on wound healing in resistance to pathogens. Annual Review of Phytopathology 27: 343-71.Bryce, J.M. (1967). The Commercial Timbers of Tanzania. Tanzania Forest Division, Utilisation Section, Moshi.Butin, H. (1995). Tree Diseases and Disorders. Causes, Biology and Control in Forest and Amenity Trees. Oxford University Press, Oxford.Butin, H. & A.L. Shigo (1981). Radial shakes and 'frost cracks' in living oak trees. USDA Forest Service, Northeastern Forest Experiment Station, Research Paper NE 478.Caspari, CO. & H. Sachsse (1990). Risschaden an Fichte-Verbreitung, Schadbild, Ursache, Auswirkungen. Forst undHolz 45: 685-688.Day, W.R. (1954). Drought cracks of conifers. Forestry Commission UK, Forest Record 26.Dick, J.H. (1969). Heartwood development and heart-rot in East African Camphorwood, Ocotea usambarensis Engl. Tanzania Silviculture Research Note No. 9.Dietrichson, J., P.A. Rognerud, O. Haveraaen & T. Skoppa (1985). Stem cracks in Norway spruce (Picea abies (L.) Karst.). Meddelelser fra Norsk Institutt for Skogforskning 38: 1-32.Junikka, L. (1994). Survey of English macroscopic bark terminology. IAWA Journal 15: 3-45.Kramer, P.J. (1983). Water Relations of Plants. Academic Press, New York.Kubler, H. (1987). Origin of frost cracks in stems of trees. Journal of Arboriculture 13: 93-97.Lundgren, B. (1978). Soil conditions and nutrient cycling under natural and plantation forests in Tanzanian highlands. Swedish University of Agriculture. Sciences. Department of Forest Soils. Reports in Ecology and Forest Soils 31.Mrema, F.A. (1997). The effect of pruning on the occurrence of stem cracks and discoloured wood in Cupressus lusitanica Miller. In L.R Mbwambo, P.L. Mwang'ingo, S.W. Masayanyika & J.A. Isango (eds), Proceedings of the Second TAFORI Workshop on Setting Forestry Research Needs and Priorities. Lutheran Uhuru Hostel, Moshi, 18-22 August, 1997. Tanzania Forestry Research Institute (TAFORI), Morogoro. Pp 94-100.Neely, D. & E.B. Himelick (1987). Freeze-crack-related measurements on Platanus × acerifolia trees. Forest Science 33: 239-244.Nsolomo, V.R. & K. Venn (1994). Forest fungal diseases of Tanzania: background and current status. Norwegian Journal of Agricultural Sciences 8: 189-201.Nsolomo, V.R. & K. Venn (1996). Decay fungi of O. usambarensis Engl. trees in the Usambara and Kilimanjaro mountain rain forests. In V.R. Nsolomo, (ed.), Fungal diseases of trees in Tanzania with emphasis on stem decay of the East African camphor tree, Ocotea usambarensis Engl. Agricultural University of Norway. Doctor Scientiarum Theses 1996: 13.Pearce, R.B. (1996). Tansley review No. 87. Antimicrobial defenses in the wood of living trees. New Phytologist 132: 203-233.Persson, A. (1994a). Stem cracks in Norway spruce in southern Scandinavia: causes and consequences. Annales des Sciences Forestieres 51: 315-327.Persson, A. (1994b). How genotype and silviculture interact in forming timber properties. Silva Fennica 28: 275-282.Renvall, P. & T. Niemelä (1993). Ocotea usambarensis and its fungal decayers in natural stands. Bulletin du Jardin Botanique National de Belgique 62: 403-414.Shigo, A.L. (1979). Tree decay: an expanded concept. United States Department of Agriculture, Agriculture Information Bulletin 419.Shigo, A.L. (1982). Tree decay. Pp 188-203 in Proceedings of the Korea-USA Joint Seminar of Forest Disease Insect Pests, 1982. Seoul, Korea.Shigo, A.L. (1983a). Tree defects: a photo guide. USDA Forest Service, Northeastern Forest Experiment Station, General Technical Report NE- 82.Shortle, W.C. (1984). Biochemical mechanisms of discolouration, decay, and compartmentalisation of decay in trees. IAWA Bulletin n.s. 5: 100-104.Struhsaker, T.T., J.M. Kasenene, J.C. Gaither Jr, N. Larsen, S. Musango, & R. Bancroft (1989). Tree mortality in Kibale Forest, Uganda: a case study of dieback in a tropical rain forest adjacent to exotic conifer plantations. Forest Ecology and Management 29: 165-185.Tate, R.L. (1986). Stem decay in street trees in New Jersey and park trees in central trees in Central Park, New York. Journal of Arboriculture 12: 73-74.Tippett, J.T., & A.L. Shigo (1981). Barrier zone formation: a mechanism of tree defense against vascular pathogens. IAWA Bulletin n.s. 2: 163-168.Trockenbrodt, M. (1990). Survey and discussion of the terminology used in bark anatomy. IAWA Bulletin n.s. 11: 141-166.Walsh, R.P.D. (1996). Climate. In P.W. Richards, R.P.D. Walsh, I.C. Baillie & P. Greig-Smith (eds), The Tropical Rain Forest: an Ecological Study (2nd edition). Cambridge University Press, Cambridge. Pp 159-205.Wilkinson, L. (1992). Systat, Inc. Intelligent software. Evanston, Illinois.