Silvopastoral Agroforestry Using

Honeylocust (Gleditsia triacanthos L.)


Andy Wilson

Springtree Agroforestry Project

268 Springtree Lane

Scottsville VA 24590, USA




    Honeylocust (Gleditsia triacanthos L.) has been identified as a temperate zone agroforestry species with potential as a pasture fodder tree.  Honeylocust trees produce pods which can provide animal feed during autumn when pasture grass production declines.  Livestock may harvest the pods directly from under


the trees, minimizing harvesting and processing costs.  Using tree protectors and/or electric fences, honeylocust orchards can be established in operating pastures.

    To test the efficiency of this agroforestry system, experimental silvopastoral honeylocust orchards have been planted in several temperate zone countries.  The various silvopastoral systems are described and preliminary findings are presented.  Economic evaluations of silvopastoral honeylocust with sheep indicate internal rates of return of 9% to 25%, depending on a variety of cost and production assumptions. 

    The combined results of existing experimental orchards should provide an improved guide to the commercial value of pastoral honeylocust in economically developing as well as developed temperate zone countries.  To help encourage an exchange of information on management practices and preliminary results, a Honeylocust Research Group has been formed.


Key Words: agroforestry, silvopastoral, honeylocust, sheep, fodder.


    Agroforestry systems integrate trees with traditional agricultural practices.  Such land use systems are economically viable when the combined value of agricultural crops and tree products is greater than the value of either product grown alone.  In tropical climates, agroforestry has often been an integral part of indigenous agriculture and has in recent times gathered wide support from international development agencies.

    Presently there are relatively few commercial temperate zone agroforestry practitioners.  Recently, ecological and financial difficulties in traditional agriculture have led to increased interest in agroforestry by the scientific community and others.  These difficulties include continuing soil erosion, falling productivity, increasing cost of petro‑chemical inputs, and environmental concerns related to the use of chemicals. Agroforestry practices can contribute to the goal of agricultural sustainability by reducing soil erosion and increasing farm profits.

    In temperate zone agroforestry, silvopastoral systems predominate [4, 21], and have relevance because of the large area devoted to grazing.  At present, temperate zone silvopastoral systems are for the most part limited to pasturing cattle and sheep in plantings with appropriately spaced timber trees.  In areas where these systems are inappropriate, alternative pastoral agroforestry designs are needed. 


Silvopastoral Honeylocust Design

    One potentially productive temperate zone agroforestry system is establishment of honeylocust orchards in operating pastures or hayfields [9, 37].  Using operating pastures allows normal cash flows from animal and hay production to continue while the honeylocust orchard matures. 

    Honeylocust trees produce pods which livestock can harvest from under the trees.  Pods drop gradually from trees over several months in autumn, and harvesting costs are nil.  The pods thus provide livestock with a complementary feed source when seasonal grass production is declining.  Although not easily quantified, additional benefits from silvopastoral honeylocust include reduction of water runoff and topsoil erosion, shade for livestock, a productive pollen and nectar source for bees, a more diversified and aesthetically pleasing pasture environment, and timber upon project termination.

    Experimentation to test the efficiency of silvopastoral honeylocust is currently taking place in several temperate zone countries (Table 1).  In addition, preliminary trials with honeylocust have also taken place in China (DR Bird, personal communication), India (F Santamour, personal communication), and Pakistan [15], but details are not available.  A majority of the current experiments have established honeylocust orchards directly in pastures with animals, and are simultaneously evaluating several cultivars as well as a variety of cultural regimes.  A variety of other system components are also being evaluated, including tree establishment techniques, tree spacing and pruning, fertilizer applications and insect control programs.

    Present research is concentrated in economically developed countries; ultimately, much of this work should have application in temperate regions of developing countries as well.  Most research is publicly financed; however, because of the need for direct evaluation in pastures with animals, cooperating farmers are playing an increasingly important role.

    When silvopastoral orchards are established in working pastures, young trees need protection from livestock browsing and rubbing.  In current experiments, three tree protection strategies are being evaluated: plastic tree shelters, electric fences, and the natural thorns presented by honeylocust. In extensive trials in the United Kingdom, plastic tree shelters have proven effective in hardwood pasture plantings with sheep and cattle [28] and are now being evaluated in honeylocust pasture orchards (C Dupraz, personal communication).  Where electric fences are used to subdivide pastures for intensive grazing and rotation, these fences provide the basis for a less expensive method of protecting trees.  While more expensive, plastic tree shelters permit a greater tree density (75 ha-1 in some trials) since trees can be planted throughout the pasture.  In developing countries where tree shelters and electric fences may prove too costly, honeylocust's natural thorns are a viable alternative.


Honeylocust Characteristics

    Honeylocust has a number of characteristics which recommend it as a silvopastoral agroforestry species [13, 26, 37].  Honeylocust is adaptable to a wide range of climates within the temperate zone.  While native to most of the Mississippi Valley, honeylocust is now found in all areas of the continental United States as well as most other temperate zone countries [5].  Honeylocust is winter hardy to temperatures below ‑34°C.  It is drought resistant (the natural range includes areas with 51 cm annual precipitation) and has good storm survival characteristics [5].

     Honeylocust thrives in a variety of soils, although it grows poorly on shallow, gravely, or heavy clay soils.  Honeylocust does well in alkaline soils, preferring a soil pH of 6.0 to 8.0, and is tolerant to acid soils and salinity.  In laboratory tests, honeylocust grown at low pH have proven to be more susceptible to aluminum toxicity than any other species tested [33].  Honeylocust may prove valuable as a bellwether species for declining pH caused by acid rain or other factors.

    Honeylocust leaves and flowers appear late in the spring, and spring frost damage is not normal.  Planting honeylocust in frost pockets, however, should be avoided.  Honeylocust naturally produce thorns (up to 45 cm) dangerous to livestock and tractor tires.  Thornless trees can be produced by budding with scion wood taken from the thornless upper branches of the desired cultivars.  While in the legume family, honeylocust does not produce nitrogen‑fixing nodules. 

    Honeylocust is polygamo‑dioecious.  For complete pollination, predominantly staminate trees must be included in honeylocust orchards.  Techniques for clonal reproduction of honeylocust using budding have been perfected in commercial nurseries.  Reproduction is also possible by grafting, root cuttings and greenwood cuttings [32].  Work with tissue culture reproduction has begun in France and in the United States (C. Borgman, personal communication).

    Honeylocust's open canopy produces a light shade, minimizing the negative effect on summer grass production.  Casual observations of field workers suggest that pasture grasses and legumes do well under honeylocust, growing right up to the trunk of the tree.  The only study of pasture grass production under a honeylocust orchard yielded inconclusive results [37, 38].  Late spring leaf‑out and early leaf drop in fall minimize shading during these seasons.  In addition, the tree's small leaflets are easily absorbed into pasture grasses during autumn leafdrop and provide an additional source of pasture fertilization. 

    Because honeylocust has been planted world‑wide as a temperate zone urban shade tree, an extensive literature exists on propagation, planting, fertilization, and disease and insect control.  A literature search in the Commonwealth Agricultural Bulletin data base for the years 1972‑1992 found 295 entries for G. triacanthos L.  This literature is available to facilitate adoption of honeylocust in agroforestry settings. 


Pod Yield and Nutrition

    Information on honeylocust pod production is incomplete, often based on studies of isolated trees and for single years, and is often anecdotal.  Information in Table 2, however,  indicates that honeylocust has good pod production under a variety of conditions.  Reports of higher and lower pod yields are available [8, 19].  Pod production is strongly biennial [26].  Pruning regimes, planting of several cultivars, and selecting improved cultivars may reduce biennial yield patterns [27].

    Studies in controlled stanchion and barnyard settings in the U.S., South Africa and France have, with one exception, produced positive results from feeding honeylocust pods to cattle, dairy cows and sheep [9,20,26,29,35].  Sheep have an ability and desire to consume the pods.  In French and South African trials, sheep ate between .75 kg and 2.0 kg of pods per day [9, 35].  Sheep can also digest between 70% and 90% of the honeylocust seed, fed in pod or shelled [9,29,30].  For sheep, 6.9% of ground pods from unselected seedlings is digestible protein (Table 3).  Cattle and swine will also consume honeylocust pods but do not digest the seeds.  No toxicity problems have been uncovered in trials to date [9, 35](G Phillips, unpublished). 


Conclusions and Discussion

    Honeylocust has often been recommended as a temperate zone agroforestry species with high potential [11,13,31,36].  Sufficient information to recommend commercial adoption, however, has been lacking.  Past studies, while giving promising results, often were discontinued before mature trees could be evaluated.  Both the number and sophistication of present studies, being conducted under a variety of soil, climate and cultural conditions, give promise that within the next decade the cumulative results of these studies will offer an improved guide to the commercial value of silvopastoral honeylocust.  An important next step will be compilation of findings from the present experiments, including information on effects of soil, climate, cultural techniques, and performance of various cultivars.  Towards this end, the author has organized the Honeylocust Research Group to improve communications and systematize data collection among professional and amateur researchers.  Following is a brief summary of progress to date, with remarks on further work needed.

Silvopastoral Honeylocust and Sheep.  Sheep appear to offer the best fit with silvopastoral honeylocust, primarily because they can digest the honeylocust seed.  Because of their smaller size sheep require less expensive tree protection.  The sheep reproduction cycle also coincides with pod production: pods can be fed appropriately in autumn to flushing ewes or weaned lambs.  For cattle and swine to receive full nutrient value, pods must be machine harvested and processed [27]. 

Tree Protection Systems.  The effectiveness and relative costs of alternative tree protection systems are being evaluated in the present trials.  Especially important in developing countries will be the effectiveness of natural thorns to protect trees.  Further work is also needed to determine the need and type of protection for mature pastoral honeylocust [16].  Glue-based paint products designed to repel livestock have been developed and are being tested in silvopastoral settings (R Sheldrick, personal communication). 

Pod Production.  While pod yields are significant (Table 2), more precise data are needed for pod production on specific cultivars and on a per hectare basis for various planting regimes.  Measuring pod production from isolated trees may give unrealistically high yields.  The development of standardized techniques for estimating yields from orchard trees will facilitate the comparison of results. 

    Fertilization regimes to increase pod yields, and the cost effectiveness of these regimes, are being evaluated.  For proper seed formation, trees with staminate flowers must be part of the honeylocust orchard.  In current trials, the percentage of staminate trees ranges from 7% to 12% [17].

Cultivar Selection.  The honeylocust displays a high degree of variability [13, 26] making it a good candidate for further cultivar selection.  Toward this end Gold at Michigan State University has established a provenance, half-sib progeny planting containing 250 sources, presently in its tenth year [13].

    Presently, 42 honeylocust cultivars selected in France, New Zealand, Australia, South Africa, and the United States are being evaluated in pasture trials with animals (Table 1).  An international exchange of seed, scion wood and grafted trees is under way.  Important cultivar selection criteria include high pod production, high pod protein, acceptability to livestock, and insect resistance.  Especially important will be selection to minimize biennial yield fluctuations.

    Feeding trials in both pastures and stanchions have begun and will play an important role in cultivar selection [9, 35].  Cultivar response to various soil types and climates will also be important.  For example, cultivar selections from the southern United States, when planted in northern regions, have produced lower yields and pods with lower sugar content, as well as having suffered winter die-back [8].

    With appropriate management practices, average annual pod production of 40 kg for 10 year old orchard trees appears possible.  A planting of 75 trees ha-1, excluding staminate trees, would yield 3,000 kg.  This output would provide 100 sheep a 1.5 kg ration for 20 days.

Pasture Management.  The effects on grass production of tree shading and competition for soil moisture and nutrients are major concerns in silvopastoral agroforestry systems.  Tree spacing and pruning are the design variables most likely to affect pasture grass production.  Spacing in the present trials ranges from 5 to 10 meters square in pasture orchards and 3 to 10 meters in fencerow plantings.  In some climates honeylocust's mottled shade may positively affect grass production (R Lance, personal communication).  Two studies have suggested that trees will affect the distribution of grass production by concentrating the grazing and resting patterns of animals around trees [38] (J Law and R Sheldrick, unpublished).  Selection of appropriate understory grass and legume species will be a priority.

     It has been suggested that using honeylocust pods as animal food will lead to the spread of thorny and unwanted seedlings in pastures [3, 19].  Most farmers presently consider honeylocust undesirable and remove them.  Five pastures with bearing honeylocust are being monitored as part of an informal survey, and some problems have been observed.  Seedlings are largely controlled by livestock grazing. 

Insect and Disease Problems.  A number of insects and diseases have been identified in the literature as potential problems affecting honeylocust growth and pod production [5, 7].  The widespread use of honeylocust as an urban shade tree has increased the number and severity of insect and disease problems encountered.  Pure stands of honeylocust in pasture orchards may be susceptible to severe disease and insect damage.  Orchard trials will indicate these effects on pod production, and subsequently the need and cost of appropriate control programs.

Alternative Scenarios.  Silvopastoral honeylocust can be used to produce several alternative outputs: pods and/or leaf fodder [2] for animal feed, fermentation of pods in ethanol production [27], pollarding for firewood [6], and trees as part of a windbreak system [3].  Currently, the most viable alternatives are windbreaks and pod production for animal feed.  Experience gained in current trials with pasture honeylocust for pod production can, however, be applied to leaf and twig fodder, energy, and firewood systems should such outputs prove to be economically viable.

Economics.  Ultimately the efficiency of silvopastoral honeylocust will be determined by the ability of this system to increase farm profits.  Preliminary economic analysis reported earlier [37], using a variety of cost and production assumptions, indicates internal rates of return of between 9% and 25% for pasture honeylocust projects.  These rates of return are based on conservative assumptions concerning pod yields and it seems likely that they will be equaled in current trials.  As these trials provide additional pod yield and nutritional data, further refinement of this analysis will be possible.  The "service benefits" enumerated above have not been included in the analysis but clearly add to the desirability of silvopastoral honeylocust.



Special thanks to Jim Allen, Gene Hunter, and Ruth Klippstein for their suggestions, and also to Greg Williams for his long-term support of this research project.




 1. Alexandra J (1982) The honey locust. In: Noel D, ed. Tree Crops: The 3rd Component pp 124-127. Subiaco, Australia, Cornucopia Press.

 2. Baertsche SR, Yokoyama MT and Hanover JW (1986) Short-rotation, hardwood tree biomass as potential ruminant feed‑-chemical composition, nylon bag ruminal degradation and ensilement of selected species. J Anim Sci 63:2028-2043

 3. Bagley WT (1988) Agroforestry and windbreaks. Agric, Ecosystems and Environ 22/23:583‑591

 4. Bandolin TH and Fisher RF (1991) Agroforestry systems in North America. Agroforestry Systems 16:95‑118

 5. Blair RM (1990) Gleditsia triacanthos L. Honeylocust. In: Burns RM and Honkala BH, Tech. Coordinators, Silvics of the North American trees Vol 2 Hardwoods, pp 358-364. US Dep Ag Handbook 654

 6. Burnett CD and Gilluly D (1988) Pollarding for multiple benefits. No J of Applied Forestry 5:148‑152

 7. Cranshaw W and Hart E (1988) Rogue's gallery: a most unwanted list of honeylocust pests. Am Nurseryman 168:68‑71

 8. Detwiler SB (1947) Notes on honeylocust.  Washington DC, US Dep Ag, Soil Conservation Service

 9. Dupraz C (1987) Un aliment concentré pour l'hiver, les gousses de Gleditsia triacanthos L.  Communication présentée aux 9èmes Rencontres du Sous‑Réseau Méditerranéen FAO pour le pâturage et la production fourragère pp. 83-87 FAO

10. Dupraz C (1993) Selection and utilization of cultivated fodder trees and shrubs in Mediterranean extensive livestock production systems. Second Intermediate Technical Report for EEC Institut National de la Récherche Agronomique, France

11. Felker P and Bandurski RS (1979) Uses and potential uses of leguminous trees for minimal energy input agriculture. Econ Botan 33:172‑189

12. Francis P (1984) Fodder trees. Farm 5:66-70.

13. Gold MA and Hanover JW (1993) Honeylocust (Gleditsia triacanthos L.), multipurpose tree for the temperate zone.  International Tree Crops Journal (In Press)

14. Hyde EAN. Pasiecznik N and Harris PJC (1990) Evaluation of multipurpose trees in southern Spain. Nitrogen Fixing Tree Res Rep 8:73‑74

15. Khan AA (1975) Chemical composition of certain trees and shrubs. The Pakistan Journal of Forestry 25:42-45

16. Kreider R (1980) My experience with honeylocust. Agroforestry Rev. 2:14‑16

17. Lennard T (1985) Use of honey locust. In: Logan LA and Radcliffe JE, eds, Fodder Trees: A Summary of Current Research in New Zealand, p 46, Christchurch. New Zealand, Crop Research Division, Dept of Sci & Indust Res

18. leRoux PL (1959) Honey locust trees: useful sources of fodder for stock. Fmg S Afr 35:20‑21

19. Moore JC (1948) The present outlook for honeylocust in the South. North Nut Grow Assoc 39th Annu Rep 19:104‑110

20. Mostert JWC and Donaldson CH (1960) Value of honeylocust as fodder tree negligible. Farming in So. Africa 35:40

21. Nair PKR (1991) State-of-the-art of agroforestry systems.  For Ecol Manage 45:5-29

22. National Academy of Science (1971) Atlas of nutritional data on United States and Canada. Washington, D.C.

23. Poynton RJ (1968) Hardy Trees for the Western Transvaal.  Bulletin No 45 So African Dep of Forestry Pretoria, South Africa

24. Putod R (1982) Les arbres fourragers: le févier. Forêt Méditerranéenne 4:33‑41

25. Roder W (1992) Experiences with tree fodders in temperate regions of Bhutan. Agroforestry Systems 17:263‑270

26. Scanlon DH (1980) A case study of honeylocust in the Tennessee Valley Region. In: Tree Crops for Energy Co‑Production on Farms, pp 21-31. US DOE Solar Energy Res Inst Symposium, Nov. 12-14, 1980, Estes Park, Co

27. Seibert MG, Williams G, Folger G and Milne T (1986) Fuel and chemical co‑production from tree crops. Biomass 9: 49‑66

28. Sibbald AR and FL Sinclair (1990) A review of agroforestry in progress in the UK. Agroforestry Abstr 3:149‑164

29. Small M (1983) Sheep can digest 90% of the whole honeylocust seeds ingested. Agroforestry Rev 3:6

30. Small M (1983) Honeylocust pods and the digestion of protein by sheep. Agroforestry Rev 4:6‑7

31. Smith JR (1950) Tree Crops: A Permanent Agriculture. New York: Derin‑Adair

32. Stoutemyer VT, O'Rourke and Steiner WW (1944) Some observations on the vegetative propagation of honeylocust. J. of For 42:32‑36

33. Sucoff E, Thornton FC and Joslin JD (1990) Sensitivity of tree seedlings to aluminum: I. honeylocust. J Environ Qual 19:163‑171

34. Tennessee Valley Authority (1984) TVA biomass fuels update II, TVA OACD‑84/3

35. Tidmarsh CEM and Roux PW (1950) Fodder Trees. Pasture Research in South Africa Progress Report No. 3. Grootfontein College of Agriculture, South Africa

36. Williams G and Merwin ML (1983) Energy‑ and soil‑conserving perennial crops for marginal land in temperate climates. In: Lockeretz W, ed, Environmentally Sound Agriculture, pp. 317-332. New York, Praeger Co

37. Wilson AA (1991) Browse agroforestry using honeylocust.  The Forestry Chron 67:232‑235

38. Zarger TG and Lutz AJ (1961) Effects of improved thornless honeylocust shade trees on pasture development.  TVA Div of Forest Relations Tech Note No 34

Table 1.  Research with silvopastoral honeylocusta

Country (State)           Year                Trees                Agroforestry system          Reference            


    Dahra                     1956‑               98 G{2}             Pasture orchard                 [24] (Dupraz, PC)


    Victoria                  1987‑               200                   Pasture orchard                 (J Alexandra, PC)

    New So. Wales       1990‑               50 G{4}             Pasture orchard                 (R Lance, PC)


    ..                            ..                     ..                      Introduction trial                 [25]


    Hérault                   1987‑               192 G{16}          Experimental orchard         [9] (C Dupraz, PC)

    Hérault                   1992‑               147 G{12}          Pasture orchard                 [10] (C Dupraz, PC)

    Hérault                   1992‑               139 G{13}          Pasture orchard                 [10] (C Dupraz, PC)

    Hérault                   ..                     .. (4 ha)             Pasture orchard                 (C Dupraz, PC)


    Nuremberg             1990‑               20 G{5}             Cultivar evaluation              (W Griesmeir, PC)


    Macedonia             1992‑               50 G{10}            Pasture orchard                 [10] (C Dupraz, PC)

    Macedonia             1992‑               50 G{10}            Pasture orchard                 [10] (C Dupraz, PC)


    Jajarkot                  1991‑               ..                      Introduction trial                 (C Edwards, PC)


    East Coast             1985‑               79 G{12}            Pasture orchard                 (RJ Hall, PC)

    So. Auckland          1958‑               ..                      Pasture orchard                 [17] (T Lennard, PC)

                                                                                 & fencerow


    Orange Free St.      1940‑1960        135                   ..                                       [20]

    Natal                      ..                     ..                      Demonstration project        (TM Maliehe, PC)

    Transvaal                1908‑?             81                     Experimental orchard         [23]

    Transvaal                1940‑1960        500                   Hayfield orchard                [35]

    Transvaal                1953‑               750                   Hayfield orchard                [18] (Grootfontein

                                                                                                                          College, PC)


    Almeria                   1988‑1989        ..                      Introduction trial                 [14]


    Alabama                 1938‑1947        95 G{2}             Hayfield orchard                [19] [26]

    Arkansas                1982‑               80 G{2}             Mowed orchard                  (K Ladd, PC) [34]

    Illinois                    1941‑               200 G{5}            Fenceline                           [16] (R Kreider, PC)

    Michigan                1982‑               ..(4.3 ha)           Provenance study              [13] (M Gold, PC)

    Minnesota              1990‑               60 G{15}            Pasture fencelines             (P Van Wazer, PC)

    Tennessee              1982‑               55 G{3}             Mowed orchard                  (H Black, PC) [34]

    Virginia                   1939‑1955        160 G{1}            Pasture orchard                 [38]

    Virginia                   1984‑               60 G{6}             Pasture fencelines             [37]


a G{n} indicates grafted trees and number of cultivars under trial.  "Pasture" indicates systems with domestic livestock.  PC = personal communication.l

Table 2.  Honeylocust pod yields (dry weight)

Location                      Kg/tree            Tree age span          Setting                    Reference

Australiaa                     25‑30                     8                         Isolated                   [12]

                                    100                         12                       Isolated                   [12]

Francea                       34                           10‑50                  Isolated                   [9]

South Africa                32                           9‑12                    Orchard                  [35]

United States

    Millwood                  33                           5‑10                    Orchard                  [26]

    Calhoun                   14                           5‑10                    Orchard                  [26]

    Millwood                  43                           9‑10                    Orchard                  [26]

    Calhoun                   16                           9‑10                    Orchard                  [26]

United States              38                           31‑34                  Isolated                   A Wilson,



a Reported single year figures have been reduced by one half to adjust for honeylocust biennial bearing pattern.





Table 3.    Nutritional composition  for sheep of ground honeylocust pods and seeds from unselected trees (dry weight basis)a

                 Crude fiber                                18.2%

                 Protein (N x 6.25)                                  10.5%

                 Digestible protein                                    6.9%

                 Digestible energy (Mcal kg-1)                 3.3

                 Metabolizable energy (Mcal kg-1)           2.7

                 Total digestible nutrients                        74.1%


a 88.4% of this pod sample was dry matter.  Values were computed from equations, not experimentally.


Source: [22]