TABLE 4. Some characteristics of four domains based on rainfall and season, as a
basis for developing weed control recommendations.

 Wet season Dry season

 High rainfall Rapid weed growth; Moderate weed growth;
 areas ready re-establish- some risks of re-estab-
 (more than ment of uprooted lishment of uprooted
 ca. 65 ins weeds and of frag- weeds and of fragments.
 [1625 mm] per ments.
 annum)
 High risk for fol- Moderate risk for
 iar-applied herbi- foliar-applied herbi-
 cides. cides.
 High risk of leach- Some risk of leaching.
 ing.

 Low rainfall Rapid, bur spor- Little weed growth;
 areas adic weed growth; very little risk of
 (less than some risk of re- re-establishment of
 ca. 65 ins establishment of uprooted weeds or of
 [1625 mm] uprooted weeds fragments.
 per annum) and of fragments.
 Moderate risk for Low risk for foliar-
 foliar-applied applied herbicides.
 herbicides.
 Moderate risk of Low risk of leaching.
 leaching



 Low rainfall-wet season Manual methods are effective ex-
cept during wet periods. Rainfall is likely to occur as short, high
intensity showers: high risk foliar-acting herbicides should be
used only when showers are not expected. Round-up can be used
in land preparation if perennials are dominant, or Fusilade if
perennial grasses are dominant. For annuals, Gramoxone should


be used. Low risk soil-acting herbicides can be used, sup-
plemented by Gramoxone sprays. Mulches can be used.
 Low rainfall-dry season Manual methods and mulching
 should be effective against the generally slight weed infestations.
 Soil-acting herbicides may be ineffective, and Round-up and
 Fusilade may not be effectively translocated unless showers
 stimulate active weed growth. Cultivations will control many
 perennials by desiccation. Gramoxone should be used only on
 growing green weeds.
 Special weed problems will require special herbicides: Reglone
 (diquat) for white heads (Parthenium hysterophorus), Talent for
 water grass (Commelina spp.) in bananas and tree crops, and
 2,4-D amine (directed) for vines.
 It is hoped in the next few years to recommend systems based
on the above guidelines.




Acknowledgements
 This paper is based on work carried out under a CARDI /USAID Project
(538-0015) from 1978-82, and currently underway as part of a CARDI/USAID
FSR/D Project (538-0099). I gratefully acknowledge the encouragement of Mr.
Calixte George, Project Manager, and Dr. Murali Rao, who commented on the
manuscript. Many other colleagues, and numerous farmers, contributed to the data
presented in this paper. My wife Gwen kindly undertook the typing.

References
 1. CARDI. 1983a. Small Farm Systems Research Project (538-0015). Final
Report 1978-82. Vol. II, CARDI, St. Augustine, Trinidad.
 2. CARDI. 1983b. Ibid. Vol. III.
 3. CARDI. 1983c. Ibid. Vol. IV.
 4. Hammerton,J.L. 1974. Problems of herbicide use in peasant farming. IPPC
Paper, Oregon State University, Corvallis, OR 97331, USA.
 5. Hammerton,J.L. 1984a. Small farm weed control in the Eastern Caribbean:
Present status and options for change. CARDI, St. Lucia.
 6. Hammerton,J.L. 1984b. Constraints to herbicide use in small farm systems.
CARDI, St. Lucia.


PROCEEDINGS of the CARIBBEAN FOOD CROPS SOCIETY-VOL. XX


136