United Nations Office on Drugs and Crime
Regional UNODC Websites



UNODC is cosponsor of the Joint United Nations Programme on HIV/AIDS - UNAIDS

Herbicidal treatments for control of Cannabis sativa L.


Materials and methods
TABLE 3 Herbicidal treatments on young and developed cannabis plants a in field plots (Field Exp. 2)


Author: Menashe HOROWITZ
Pages: 75 to 83
Creation Date: 1977/01/01

Herbicidal treatments for control of Cannabis sativa L.

Division of Weed Research, Agricultural Research Organization, Newe Ya'ar Experiment Station, P.O. Haifa, Israel


In order to test herbicides for the destruction of illicit stands of cannabis ( Cannabis sativa L.) a series of commercially available herbicides were sprayed on glasshouse-grown plants having 2 to 6 leaves. The following herbicides caused complete kill or severe injury to cannabis plants: (a) herbicides with root and foliage activity - ametryn, atrazine, metribuzin, prometryn, terbutryne, diuron, fluometuron, linuron, methabenzthiazuron, phenobenzuron, ethofumesate, karbutilate, methazole and oxadiazon; and (b) foliar-acting herbicides with brief or no soil persistence - amitrole, bentazon, 2,4-D, diquat + paraquat, glyphosate and phenmedipham.

In field experiments herbicides of the latter group, and ioxynil, metribuzin, and a MSMA-cacodylate mixture, caused death or severe damage to young cannabis plants. Glyphosate, ioxynil and bentazon destroyed developed cannabis plants.

In glasshouse and field experiments the following herbicides applied to young cannabis plants caused marked deformations of stems, leaves and/or inflorescences: barban, butralin, dalapon, difenzoquat, dinitramine, diphenamid, IPC, napropamide, penoxalin, trifluralin, and U-27267.


The spread of drug abuse is of great concern to all civilized countries. One means of combating the drug problem is to eliminate the illicit cultivation of narcotic plants. The present work was undertaken as part of an international effort to provide the appropriate authorities with methods for effective eradication of illicit stands of cannabis ( Cannabis sativa L.).

The objective of the research was to screen for herbicides which would destroy or heavily damage cannabis plants - after their presence had been detected and found undesirable. Another line of screening was based on data by Gentner (personal communication), reporting that certain herbicides could induce abnormal flowering on Papaver somniferum L.; it was thought that compounds capable of impairing the normal development of cannabis plants could also constitute a valuable means of control.

Several prerequisites limit the number of chemicals utilizable for these purposes. The herbicide must be active in post-emergence application and have a minimum impact on the environment, i.e., negligible drift towards surrounding crops and limited persistence in soil to avoid residual damage to subsequent crops in the treated field.

In a first series of experiments, commercially available herbicides belonging to different groups and having various types of foliar activity, were sprayed on plants grown in containers in the glasshouse; their response was observed until flowering. In the second stage, promising foliar-active herbicides with limited residual effect were applied on cannabis plants grown in the field.

Materials and methods

Glasshouse experiments. Seeds of two varieties obtained from the U.S. Department of Agriculture (U.S.D.A., Plant introduction numbers 377927 and 377929) were sown in 150 ml plastic cups with drainage holes at the bottom, filled with Newe Ya'ar clay soil. Altogether, 49 herbicides were tested in nine series, prepared at approximately monthly intervals. Plants had between two and six leaves at spraying time; for each series plants were chosen to be as uniform as possible. The herbicides were applied with an automatic sprayer in 200 litre of water per ha. Each compound was sprayed at two doses, and an untreated control was included in each series. Treatment of each variety was repeated five times.

The plants were placed in a glasshouse before and after application. During the period of work monthly minimal temperatures varied between 6.5 °C (January) to 19.9 °C (July); maximal temperatures varied from 33.8 °C (January) to 41.8 °C (July). The treated plants were watered for the first time two days after spraying; otherwise they were watered as necessary. In most cases treated plants were kept under observation until flowering. Observations were visual, using a rating scale of 0 to 10 for the degree of injury (0 = no injury, 10 = complete kill).

Field experiments. Seeds of another variety obtained from the U.S.D.A. (Plant introduction number 377935, of Turkish origin) were used in all field trials. Three experiments were carried out in 1974 and one in 1975 (Exp. 4), all on Newe Ya'ar clay soil. In each experiment several rows of 100 m length were sown at 1- m distance. Herbicides were sprayed in 200 litre water per ha with a portable sprayer fitted with a Tee-Jet nozzle on a 1.5 m boom. Observations were made visually, as in the glasshouse.

Exp. 1. - Plants about 10 cm tall were sprayed on 1 July 1974.

Exp. 2. - Two adjacent plots were sown 2 weeks apart and sprayed on 1 August 1974, when young plants were 10-20 cm tall and older plants were about 80 cm, just starting flowering.

Exp. 3. - As in Exp. 2, there were two plots, will younger plants 15-25 cm tall and developed plants 50-80 cm tall, both sprayed on 12 September 1974. All treatments in Exp. 1, 2 and 3 were repeated four times. Absolute minimum and maximum temperatures during the 1974 period of experimentation were 15.2 and 36.0 °C in Exp. 1, 14.7 and 36.0 °C in Exp. 2, 12.7 and 35.3 °C in Exp. 3.

Exp. 4. - Herbicides were applied on plants which were in full flowering, about 60 cm tall, on 14 October 1975. Each treatment was repeated three times. Growing conditions were severe and plants suffered from insufficient soil moisture. Absolute minimum and maximum temperatures during the period of experimentation were 12.6 and 35.5 °C; in the week following the treatment there was a spell of hot and dry ("sharav") weather.


Forty-nine herbicides were sprayed on cannabis plants of two varieties grown in containers and having two to six leaves. There was no marked difference in response between the two varieties. Each herbicide was applied at two doses; the lower one was, in general, the dose recommended in common Israeli field practice. The treatment which caused complete kill (grade 10 of visual assessment), severe injury (grade 8-9), or sub-lethal deformations, are listed in table 1.

Among the 20 herbicides which killed or severely injured cannabis plants there are five triazines (ametryn, atrazine, metribuzin, prometryn and terbutryn), five substituted ureas (diuron, fluometuron, linuron, methabenzthiazuron and phenobenzuron), and four belonging to other chemical groups (ethofumesate, karbutilate, methazole and oxadiazon). These compounds are known mostly for their effective pre-emergence and early post-emergence activity. A few days after spraying, treated cannabis plants showed an appreciable check on growth as compared with the controls; narcotic zones developed on the leaves and within 1 to 2 weeks the plants collapsed.

TABLE 1 - Herbicidal treatments which killed or injured cannabis plants a in glasshouse trials.

Complete kill

Severe injury


Ametryn 250 +
Amitrole TL 2500
Barban 750
Atrazine 250 +
2,4-D 660
Butralin 2000 +
Bentazon 2000 +
Dalapon 4250
+ paraquat 200 +
Difenzoquat 1600
+ paraquat 400 +
Ethofumesate 4000
Dinitramine 960 +
Diuron 800 +
Diphenamid 4800
Fluometuron 800 +
thiazuron 1400 +
IPC 7500
Glyphosate 820 +
Phenmedipham 825 +
Penoxalin 660 +
Karbutilate 2000 +
Phenobenzuron 1250 +
Napropamide 2500 +
Linuron 750 +
Trifluralin 960 +
Methazole 1875 +
U-27267 3000
thiazuron 2800
Metribuzin 350 +
Oxadiazon 500 +
Phenobenzuron 2500
Prometryn 250 +
Terbutryn 250 +

a Foliar spraying on plants with 2 to 6 leaves.

b Each compound was applied at two doses; + indicates that the effect was produced by the lower tested dose.

The following herbicides, effective against cannabis, are typical foliar-acting compounds: diquat + paraquat mixture, amitrole, bentazone, 2,4-D, glyphosate and phenmedipham. Diquat + paraquat acted rapidly, and wilting already appeared several hours after the application; at the higher dose all plants died within 3 days. The other compounds, except 2,4-D, produced symptoms and achieved their final effect much more slowly. Bentazon killed the plants after one week; with glyphosate the plants lost their vigour, yellowed after several days and collapsed about two weeks later. Amitrole TL, 2,4-D, and phenmedipham caused severe injury but only partial kill at the rates tested. On amitrole-treated plants, typical discolouration was apparent after 3-4 days, followed by gradual wilting of leaves. With 2,4-D, "hormone" injury was observed already one day after spraying: upward bending leaves and twisting of stems; growth deformation and leaf wilting became more severe with time, although most damaged plants remained vegetatively alive. Heavy damage from phenmedipham appeared approximately one week after application.

Several treatments induced vegetative or reproductive abnormalities without actually killing or severely damaging cannabis plants. In most of these treatments, the elongation of stems and the differentiation of inflorescences were delayed and reduced; percentage of normal inflorescences was markedly reduced with butralin, dinitramine, penoxalin, U-27267, and strongly reduced with barban, napropamide and trifluralin. Leaves and inflorescences had, in many cases, an abnormal appearance as compared with the untreated control. For instance: trifluralin-treated plants had shorter and thicker stems; upper leaves were narrower; in male plants panicles were thinner, and female inflorescences appeared more dense; differentiation into male or female inflorescences was often incomplete. Butralin, dinitramine and penoxalin, which belong to the group of dinitroaniline herbicides, like trifluralin, also produced thicker stems and abnormally looking leaves and inflorescences. Deformation of male inflorescences, and to a lesser extent of female inflorescences, occurred also among plants treated with diphenamid, napropamide and U-27267 (amide herbicides), and with barban, dalapon, difenzoquat and IPC (grasskillers of various chemical compositions). In addition to abnormal development, some of these treatments caused scorching (dalapon, difenzoquat, dinitramine) and discolouration (difenzoquat, napropamide, U-27267). However, most plants sprayed with these deformation-inducing compounds remained alive; in a visual assessment rating total injury from 0 (normal) to 10 (killed), the grading of plants treated with the higher dose of difenzoquat and trifluralin was 7, 6 for napropamide; and 5 to 4 for the other herbicides of this group.

The following herbicides showed minimal activity in glasshouse experiments: alachlor, benzthiazuron, benefin, butylate, chloroxuron, cycloate, E.P.T.C., molinate, nitralin, pebulate, perfluidone, pronamide, T.F.P. and vernolate. The foliar-active herbicide which was best tolerated by cannabis was nitrofen, even at 5 kg/ha; this compound could be considered for use to control weeds in licit stands of cannabis.


Herbicides which produced strong foliar activity in the glasshouse trials, with minimal persistance in soil, were tried in field experiments. Ioxynil and arsonates were included in these experiments although not tested in the glasshouse, because of other relevant research data.

In Field Experiment 1 (table 2), carried out on plants about 10 cm tall, the mixture of diquat + paraquat caused severe scorching of the foliage already one day after spraying, and complete kill of the plants a few days thereafter.

TABLE 2 - Effect of herbicides on young cannabis plants a in field plots one and two weeks after spraying (Field Exp. 1)


Injury rating (b)


Dose a.i. kg/ha

1 week after spraying

2 week after spraying

2.000 9.5 10.0
Diquat + paraquat
0.400 10.0 10.0
0.350 9.5 10.0
M.S.M.A. + cacodylate
1.900 6.6 8.6
0.660 8.6 8.2
  0 1.2

a.) Foliar spraying on plants about 10 cm tall.

b.) Visual rating from 0 (= normal) to 10 (= complete kill).

Bentazon and metribuzin acted more slowly, but also achieved complete kill - in approximately 10 days. Phenmedipham and the mixture MSMA-cacodylate produced heavy damage, but many plants remained alive.

TABLE 3 Herbicidal treatments on young and developed cannabis plants a in field plots (Field Exp. 2)


Weeks after spraying small plants

Weeks after spraying developed plants









2.000 4.4 7.6 10.0 3.0 5.8 6.6
Amitrole TL
2.500 6.0 9.6 10.0 3.2 6.6 6.8
0.820 9.2 10.0 10.0 5.6 10.0 10.0
0.140 8.8 9.0 8.2 5.0 4.0 3.0
0.280 9.8 10.0 10.0 4.6 5.6 4.6
1.000 10.0 10.0 10.0 6.2 5.8 3.6
1.000 5.0 4.0 3.8 3.2 0.6 0.4
0.2 0.6 0.4 0 0 0  

a.) Foliar spraying on young plants 10-20 cm tall and on developed plants about 80 cm tall.

b.) Visual rating from 0 (= normal) to 10 (= complete kill).

In Field Experiment 2 (table 3) herbicides were sprayed on plants which had just started flowering and were about 80 cm ' tall, and on younger plants, about 10 to 20 cm tall. Metribuzin, which killed small cannabis plants in the previous experiment at 350 g/ha, was tested again because of its considerable herbicidal activity. In the present trial, young plants were killed by 280 g/ha and severely damaged by 140 g/ha. Developed plants, however, were only scorched with these two doses, and partially recovered. Amitrole and anitrol + thiocyanate (TL) induced a slow discolouration of the foliage and both killed young plants after 3-4 weeks, but were insufficient on developed plants. Ioxynil caused rapid death of small plants, while developed plants recovered from the initial scorching. Bromosynil was much less active than ioxynil and only plant parts directly hit by the spray were damaged.

Glyphosate at 820 g/ha was the only herbicide which killed both young and developed cannabis plants within 2 weeks.

In Field Experiment 3 were tested herbicides which induced vegetative and reproductive abnormalities in previous glasshouse trials. Small and developed cannabis plants were sprayed with seven herbicides: barban, 0.75 kg/ha; butralin, 4.0 kg/ha; dalapon, 5.1 kg/ha; difenzoquat, 1.6 kg/ha; dinitramine, 1.92 kg/ha; IPC, 7.5 kg/ha; and trifluralin, 1.92 kg/ha. In addition to causing deformative development, several compounds had a direct scorching effect on plants. Dinitramine was the most damaging herbicide, since treated plants were severely scorched. Almost all young plants were killed. On developed plants, effects of dinitramine were slower to appear and were distributed irregularly; plants which did not wilt, bore abnormal leaves and inflorescences. All other treatments damaged but did not kill plants.

In Field Experiment 4 several herbicides included in previous experiments were tested again, but later than previously, on plants in full flowering and under dry conditions. Assessment was made mainly on female plants, since male plants dried out soon after flowering. The diquat + paraquat mixture at 400 and at 800 g/ha scorched severely the peripheral leaves, without penetrating into the more protected parts of the plant. This was also the case of metribuzin at 175 g/ha, although at 350 g/ha most leaves and some of the inflorescences wilted. Glyphosate at 820 g/ha caused slow wilting of external leaves and yellowing of inflorescences but, unlike in the previous experiment, failed to kill the plants. Under the particular conditions of this experiment, ioxynil at 1250 g/ha and bentazon at 2500 g/ha were the most efficient compounds, producing almost complete wilt of leaves and inflorescences.

Dalapon caused scorching and growth check on both young and developed plants; leaves had a crumpled appearance and some of the inflorescences failed to develop. Difenzoquat had only a slight effect on young cannabis, but, surprisingly, leaves of developed plants were severely scorched and subsequently wilted. Barban caused some leaf scorching and reduction in growth. Butralin and trifluralin induced deformations similar to those due to dinitramine but less severe and not accompanied by leaf scorching. In this experiment IPC had no appreciable effect on young or developed plants.


In the glasshouse trials, 20 herbicides (out of 49 tested) killed young cannabis plants or injured them severely, probably precluding any further use of their leaves and inflorescences for normal narcotic production (table 1). Most of these herbicides - five triazines, five substituted ureas and four of other groups - act primarily through root uptake. The observed foliar effects are due, in some cases, to direct foliar absorption, or are caused by root uptake and upward translocation. In common practice, these compounds are sprayed before weed emergence and often also in early post-emergence. For eventual application of these herbicides on developed cannabis plants, two points must be taken into consideration: ( a) their foliar activity is generally limited to young and small plants - altough it can be extended markedly by increasing the dose or adding surfactants; ( b) they have an appreciable persistence in soil, which varies directly with the dose applied.

Six foliar-active herbicides were particularly effective in glasshouse trials: amitrole, bentazon, 2,4-D, diquat + paraquat, glyphosate and phenmedipham. These compounds are commonly sprayed on the foliage of weeds. Their main sites of uptake and of action are in the leaves, and their residual activity in soil is nil or brief. These compounds, with the exception of 2,4-D, were tested subsequently under field conditions.

2,4-D and other derivates of the phenoxy group are known to cause severe deformation, leading generally to death, on many cultivated plant species. However, they have a serious environmental impact, as most formulations are quite volatile (the 2,4-D used in glasshouse trials was a relatively low volatile acid formulation) and their drift is capable, even in very low concentration, of inducing highly conspicuous abnormalities to surrounding vegetation.

In the field experiments were also included two compounds not tested in the glasshouse trials: ioxynil, a foliar-active herbicide with no soil persistence, related to bromoxynil, and a mixture of arsonates, MSMA + cacodylic acid and salt, which has a strong scorching effect under high temperatures. Among herbicides having both soil and foliar activity, only metribuzin was tested in the field, because of its strong phytotoxicity at rates which disappear rapidly in soil.

On young cannabis plants grown outdoors, all tested compounds, except bromoxynil, caused complete destruction or at least severe, and probably irreversible, injury (tables 2 and 3). It is remarkable that in most cases results obtained in the field were similar to those on glasshouse-grown plants of the same age. However, on larger plants the herbicidal effect was much less than on young plants. In one experiment, only glyphosate was effective. In another experiment, when ioxynil and bentazon were sprayed on plants in full flower and under particularly dry conditions, they produced heavy damage, whereas glyphosate was insufficiently active. These results confirm a known principle of weed control, that the size, stage of development and physiological condition of plants affect considerably the outcome of treatments. As indicated previously, by increasing the dose and adding suitable surfactants it is often possible to increase the foliar activity, but this has to be tried under local conditions. Obviously, control measures should be applied at as early a stage of growth as possible.

Another approach to elimination of narcotic production from cannabis, short of killing the plant, is to cause marked deformations of its foliage and inflorescences, which are likely to prevent their utilization or, at least, to reduce their commercial value. In his work on opium poppy ( Papaver somniferum L.), Gentner found that herbicides such as O.I.P.C., barban or trifluralin, applied on young plants, induced abnormal flowering (personal communication). Along the same ,line, several carbamates, dinitroanilines, and herbicides of other groups were sprayed on cannabis plants in the glasshouse and in the field. The herbicides chosen for this purpose had no known foliar activity on dicotyledons and were chosen on that basis in order to avoid interference with eventual deformations. The following herbicides induced deformations of stems leaves and/or inflorescences: four herbicides which are primarily grasskillers-barban, delapon, difenzoquat and I.P.C.; four dinitroanilines - butralin, dinitramine, penoxalin and trifluralin; and three amides - diphenamid, napropamide and U-27267. It is noteworthy that herbicides of the two latter groups are usually sprayed on the soil before weed emergence, and are considered to act on germination and through root uptake. It is possible that the commercial formulation used is responsible for their foliar uptake and activity; this can probably explain the strong scorching caused by dinitramine. Visually, the most severe abnormalities were produced by dalapon, dinitramine, difenzoquat, napropamide and trifluralin. The influence of these herbicides on the drug content was not investigated in this work.

Data on persistence in soil and selectivity to various crops of herbicides which killed or injured cannabis plants are presented in the appendix 2. These indications can be used as a guide for choosing the herbicide which will produce the minimum environmental disturbance.


This research was conducted under contract with the U.S. Agriculture Dept., which provided also the seeds for experimental cannabis growing. I wish to thank Dr. W. A. Gentner, A.R.S., U.S.D.A., Beltsville, Maryland, USA, who was sponsor of the research programme, for his advice and help.

The experimental work was made at the Newe Ya'ar Experiment Station on cannabis grown according to special authorization from the Israeli Ministry of Health. I wish to express my appreciation to Ing. Agr. G. Herzlinger and the technical staff of the Division of Weed Research, for their assistance.

back to top