Significance of additives to enhance the acceptance of poison bait in poultry rodents of Haripur, Khyber Pakhtunkhwa, Pakistan

Rodent infestation on poultry farms incurs heavy economic losses to this industry by causing feed loss and disease introduction. Development and continuous improvement of rodents control techniques are vital to minimize and control the damages caused by rodents. Here, we test the feed preference of rodents for locally available and palatable food grains viz. millet (whole), wheat (cracked) and rice (broken) and taste additives namely whole egg (5%), eggshell (5%), peanut cracked (5%) and yeast (2%) that were offered mixed in millet-wheat (50:50 by wt.) bait. We tested the preferences of different food additives through a process of feed choice mechanism. We applied two different techniques to compare the preference of mixed feed baits, these techniques included no-choice with multiple choice feeding tests and paired choice with multiple choices feeding tests. The results indicated that consumption of bait with added whole egg was significantly higher (p > 0.05). Further test for its effectiveness as a carrier for rodenticides revealed 56%, 82% and 92%, reduction in rodent activities with zinc phosphide (2%), coumatetralyl (0.0375%) and Brodifacoum (0.005%) respectively. Our results point to a need on continuous improvement of feed baits by using different combinations to effectively control the rodent infestation.


Introduction
Commercial poultry farming is scaling up rapidly at global level to fulfill the food requirement of the increasing human population.Several farmers in developing countries are taking up poultry to increase their income.The industry is expanding at 2-3% per year and is expected to continue to grow.However, this industry is not immune to the modern world problems and is severely affected by rodent infestations globally [1].Rodent pests cause heavy losses to poultry farms globally [2,3].The precise estimates of these losses caused by rodent pests are not easy to document as they affect at qualitative and quantitative level [4].
Several trials have been conducted to assess the food preferences of rodents to prepare the effective poison baits.The trials in captivity suggested that millet and rice were preferred by M. domesticus, R. rattus and R. norvegicus [5][6][7]; M. musculus preferred canaries, seeds, oats and wheat compared to soft wheat, corn, sorghum, sunflower seeds and barns + 5% molasses [5,[8][9][10] and corn were preferred by R. Exulans and, M. musculus [11,12].A trial in London found that a mixture of whole wheat, canary seeds, wheat and pinhead, oat meal is preferred by M. domesticus [10].The Canary seeds, the wheat, and the rice attracted M. domesticus in Australia [5].Whereas in Egypt, M. musculus chose barley, over, sorghum, seeds, sunflower, and bran with, added 5% molasses [8,9].There are limited local case studies to assess the food preference of local rodent species in Pakistan.In one study locally available taste additives namely, broken peanut (5%), yeast (2%), fish meal (5%), egg shell powder (5%), carbon disulfide (CS2) 30ppm and jaggery (5%) were provided to poultry rodents by mixing in millet-wheat (1:1) bait.This study revealed the order of choice of different bait formulations by different rodent species where egg, shell bait was the most preferred whereas the peanut bait was found to be least preferred [13].
Most of the previous studies used taste additives at, 2% (w t /wt) to, a bait, base composed, of, equivalent, amounts, of, broken, rice, and wheat, flour.Taste additive1such as eggs (3%) are known to make bait more palatable to commensal, (R. rattus) and field rodent (N.indica and, B. bengalensis) [14].In another study in, India 2% egg albumin and egg shell, powder in cereal bait enhanced the acceptance and efficacy of 2% zinc phosphide against R. rattus trapped in poultry farm conditions [15].In Pakistan, 5% egg, shell, grain bait was evaluated for its efficacy as a carrier for zinc phosphide (2%) and coumatetralyl rodenticides (0.0375%).The decline in rodent activity after zinc phosphide and coumatetralyl, treatment was 70% and 82% respectively, [13].
However, prolonged use of same bait combination at a particular location decreases the efficacy of that bait leading to an increased rodent infestation.To keep baits attractive and affective for rodent control a continuous research process should be carried out to develop multiple poison bait combination for effective rodent control [16].The major objective of this, study, was to establish preferences among millet, wheat, and rice Furthermore, we tested the additions of some substance such as whole egg, eggshell, yeast, and peanut contribute to increase the consumption of these grains, and whether these additives promote the consumption of rodenticide baits.

Materials and methods
Advanced Studies Research Board ethical committee of University of Haripur, KPK, Pakistan approved all experiments.All experiments were performed in accordance with relevant guidelines and regulations.
We conducted this study in the poultry farms of Haripur in the year 2018 and 2019.The study area covered 3465.6 km 2 (33 0 36'N, 73 0 04'E).Most of the poultry farms were located in areas surrounded by harvested wheat, corn, sorghum and groundnut crops.
The design of bait stations plays an important role in bait consumption of commensal rodents.Our locally made wooden box bait stations are 30 cm long 19 cm wide and 9.5 cm high with two entrances (5 cm in diameter) at opposite ends of the box [13].There was an internal 6 cm long bait positioning channel and a lockable lid at the bait station (Fig 1).To determine the spillage and keep the baits out of the approach of the chickens, the grains were offered in bait boxes [17] providing bait inside bait boxes is an efficient form of rodent management [18].
We selected three types of locally available grains that among the most palatable to rodents [13,[19][20][21] to further assess their palatability and acceptance by the rodent populations in poultry farms in our study area.We tested wheat (Triticum aestivum) in cracked form, rice (Oryzasativa) in broken form and whole millet seed (Pennisetum typhoideum) in no-choice and multiple-choice feeding tests to determine the most preferred grain.

No choice feeding
We offered three grains viz.millet, wheat and rice.Wheat grains were ground/cracked to a particle size similar to that of millet grain, while rice (in 1broken form).All three grains were cleaned of debris and dust before use.In no-choice test at each bait station each grain was offered in one bait box for 17 days at the three selected poultry farms.Bait boxes were placed at appropriate where there was evidence of movement/activity of rodents.Initially, we offered 100 g at each bait station, but the amount was subsequently adjusted to ensure a surplus amount of feed for the rodents.Bait materials were replenished twice mean while consumption and spillage were measured by using electronic digital balance.On each visit position of bait boxes were changed to avoid the place preference [13,22].The grain consumption was estimated by taking the difference in amount of bait placed and the amount of bait left over in each bait station on the day of observation.Numbers of bait stations varied in accordance with the size of the poultry farms and apparent signs of the rodent activities.
There were eight bait stations at the first farm/replicate, six at second and third farm/replicate.The bait stations were located inside the poultry farm buildings (feed store and poultry shed) and even at open areas outside the buildings, if 1necessary.However, it was preferred to place the bait boxes inside the buildings to avoid chances of theft, exposure to non-target animals and extreme weather conditions.

Multiple choice feeding test
Multiple-choice feeding studies were conducted on the same poultry farms approximately four weeks after the no-choice trials.Millet and wheat were delivered simultaneously in sets of two bait boxes located at the appropriate locations at the three poultry farms in this feed test.We used eight bait stations at farms one and two while six at the third in pairs of four and three, all of the bait boxes were placed in pairs.

Additives
We picked four locally accessible bait additives which have proven to have some efficacy, i.e., egg shell powder, peanut, and whole egg [13,23] and yeast [13,24].These additives were used at 2% and 5% concentrations to test their potential.Based on the no-choice experiment and multiple-choice feeding test results, millet (whole) and wheat (cracked) were used as bait.

Paired choice feeding test
We included four poultry farms to evaluate a suitable bait additive for attraction and enhancement of bait intake.All baits were prepared by adding and mixing thoroughly a measured quantity (2% or 5% w/w) of the additives to the bait.The feeding tests were performed simultaneously at four selected poultry farms using a paired choice procedure offering plain bait (mix of millet and wheat) in one box and additive bait in the second.We placed bait boxes at locations where there were signs of activity of the rodents.The number of bait stations ranged from six to eight.

Multiple-choice feeding tests
This experiment was performed at three poultry farms to further refine the results of pairedchoice feeding tests.A mix of millet and wheat with additives were offered simultaneously in separate bait boxes placed together on one bait station.There were four sets of bait stations at each poultry farm to place five bait boxes per station.

Rodenticides
The practical field efficacy of the most palatable candidate bait ingredients was tested by incorporating commonly used rodenticides.Acute rodenticide (zinc phosphide) and anticoagulant (coumatetralyl 0.0375% and brodifacoum) rodenticide.

Efficacy of zinc phosphide
Zinc phosphide bait was experimented with at three poultry farms using pre-baiting for six days to mitigate neophobia and increase the chance of rodents consuming a lethal dose [13,25].Zinc phosphide bait was then offered for three consecutive days in eight bait boxes at farm one, seven at farm two and ten bait stations at farm three.To assess pre-treatment rodent activities: we placed vinyl tiles (15 x 15 cm) with chalk powder spread on their surfaces in feed/egg stores and poultry sheds for three consecutive nights.We observed these tiles every morning for rodent footprints.Post-treatment rodent activity was recorded immediately after the removal of the bait boxes.Change in ratio used to estimate the success of the rodenticide bait.

Efficacy of coumtetralyl/raccumin (0.0375%)
An additional set of three poultry farms was selected for evaluation of bait containing 0.0375% coumatetraly using 6 bait boxes at the first farm, 8 at second farm and 11 at third farm.Pre-treatment rodent activity was recorded using tracking tiles as before.The bait was placed for 11 consecutive days, and then tracking tiles were placed again for three nights.

Efficacy of brodifacoum (0.005%)
For the assessment of bait containing 0.005% brodifacoum, another set of three poultry farms were selected using 8 bait boxes at the first farm, 10 at the second farm and 7 at the third farm.Using tracking tiles as before, pre-treatment rodent activity was recorded.For 11consecutive days, the baits were placed and afterward tracking tiles were set again for three nights.

Statistical analysis
➢ We employed ANOVA to assess the effect of farms, days of feeding and grain type on bait consumption for no choice and multiple choice feeding tests of plain baits and multiple choice feeding tests of additives.To analyze non independent data, because the feeding tests were conducted in the same farms, it was conducted as repeated measure analysis.
➢ The t test was used to compare the mean consumption of bait with additives versus plain bait in the paired choice test.
➢ Rodent activity on tracking tiles before and after treatment was compared using paired, two tailed t tests.Changes in rodent activity indices were assessed using the method of [13,26].

No-choice test
Examination of the intake of three palatable grains (millet in whole form and wheat and rice in cracked forms) over the 2.5 weeks (17 days) period revealed that farms, days of feeding and grain types all had a signicant effect on bait take (Tables 1 & 2: F = 7.414, df = 2, P = 0.004).Bait consumption at farm 3 was higher than at farms 1 and 2. Rodents ate significantly more millet and wheat than rice (Table 3).Neophobia (avoiding a new thing/object in the environment) is often displayed by rodent species and explains the low consumption of grain during the initial four days of millet presentation in the no choice test.The rst signs of rodent visits to bait boxes were recorded on day 5 of baiting, whereas more than 65% of the bait boxes showed rodent activity on day 11 and 70% by day 17, with a similar level of activity for the remainder of the trial.It was reected in a gradual but signicant increase in millet consumption with time, and we observed a similar pattern in the consumption of the other two grains (wheat and rice).The difference in consumption of the three grains during the study was presumably inuenced by taste, odour and palatability of the individual grain type and differences between farms and availability of other food types.

Multiple-choice feeding test
The results of the multiple choice test conrmed those of the no choice tests, with millet again the most preferred grain to poultry rats (Tables 4 & 5: F1 = 6.702, df = 1, p = 0.024).Days of feeding and grain types signicantly inuenced bait consumption.The consumption of millet and wheat did not differ significantly.
These tests were performed on the same farms as the no choice tests, and after a 2-3 weeks interval rodents did not show 1initial neophobia towards the bait stations, with 60% of bait stations showing rodent activity on day 5 of the trial.Whereas in the no choice trial the amount of each of the grains eaten showed a gradual increase in food consumption.

Paired choice feeding test
The results of the paired-choice feed tests (bait combined with additive versus bait without additive) are listed in the following subsections.

Whole egg
Comparison of total grain ingestions showed that the ingestion of whole egg bait was significantly higher than that of plain bait (Table 6; t =-4.1, df = 6, p = 0.001).After some initial hesitation, both bait boxes showed signs of rodent activity, although a marked preference for the additive bait over the plain bait was recorded, consumption of plain bait uctuated while that of additive bait increased linearly (Table 6).The mean consumption of the bait with egg additive (18.47 ± 1.85g) was higher than that of plain bait (8.2 ± 1.91 g).

Yeast
Cumulative consumption of bait added with 2% yeast powder was signicantly higher than that of the plain bait (Table 7: t = -13.823,df = 6, p = 0.009).

Peanuts
Bait comprising mixed millet and wheat with 5% cracked peanuts added was eaten in large quantity than the plain bait mixture (Table 8; t = -4.05,df = 6, p= 0.007).On the 3 rd day all 5 bait boxes (for plain & additive baits) showed signs of rodent activity.Consumption of both bait forms fluctuated with no consistent trend over time (Table 8).

Egg shell
A comparison of the total grain intake showed that bait consumption was significantly higher with added egg shells than the plain bait (Table 9: t = -1.95,df = 6, p =0.007).All the bait boxes showed signs of rodent activity after some initial hesitation, although a marked preference was observed for the additive bait over the plain bait (Table 9).The plain bait consumption fluctuated, while additive bait consumption rose linearly.The mean consumption of the bait with egg additive (16.05 ± 0.82 g) was higher than the plain bait (11.47 ± 1.39 g).

Multiple choice feeding test
When all baits with additives and plain bait were tested together consumption of bait with 5% whole egg was the highest (Table 10); there was also signicant differences between the different baits (F = 37.774, df = 4, p < 0.001).Farms, days of feeding and additive types, had a signicant inuence on bait take (Table 11).
Comparison of bait types using Duncan's Multiple Range test suggested the following order of preference; whole egg bait > egg shell bait > yeast bait > peanut bait > plain bait.Consumption of whole egg bait was signicantly higher than egg shell bait and consumption of both of these formulations was signicantly higher than the other three bait formulations.All 15 bait boxes containing egg shell and yeast baits showed positive signs of rodent activity within three days of their placement (Table 10) while with plain bait it took 10 days.The rest of the bait types showed varying attractiveness to the rodents.

Efficacy of the most palatable bait with 2% zinc phosphide
Unlike the previous feeding1tests the rodent populations1on1the three poultry1farms1visited the bait boxes on the rst day1of their1placement.Consumption of plain 1 bait (Table 12) 1during the first three days of pre-baiting increased from 33 g per bait box per night, 41 g per bait box per night on day 4. When the rodenticide1bait was offered on the 7th day of the test there1 was a reduction in bait1consumption to 20±13.22g/night/baitbox.Afterwards there 1 was a 1gradual1decrease in consumption of poison1bait presumably due to the 1rodenticide-related The pre-treatment and post-treatment rodent monitoring showed a significant reduction (t = 12.8, df = 2, P = 0.006) in rodent activity in the post-treatment period compared to the pre-treatment period.The overall rodent activity reduction could be due to 2% zinc (Table 13) phosphide (mixed with wheat-millet & whole egg) treatment was estimated at 56%.

Efficacy of the most palatable bait as carrier of 0.0375% coumatetralyl
Mean daily consumption of coumatetralyl bait for the rst eleven days was 22 ± 3 g/positive bait 1box/night (Table 14).Afterwards, the intake of coumatetralyl bait started to decrease, and only 112 g ± 1 g/ positive bait box/night got consumed on the 14th night of treatment.
We found the carcasses of rodents with symptoms consistent with anticoagulant poisoning.Three dead rodent species, i.e., Mus musculus, Rattus rattus and Rattus norvegicus, were retrieved from three poultry farms.Four dead rodents were found on the 4th night of exposure, six after the 15th night of exposure and increasing numbers up to night 11 (Table 15).After eleven nights of exposure, there was a decrease in the intake of poison bait and the number of dead rodents.The observation of pre-treatment and post-treatment period revealed a significant decrease (t = 9.5, df = 12, P = 0.03) in the post-treatment duration of rodent activity.The average decline in rodent activity (Table 15) was 82% due to ingestion of the 0.0375% coumatetralyl bait.

Efficacy of the most palatable bait as carrier of 0.005% brodifacoum
A mean daily consumption of brodifacoum bait for nine days was 22±3 g/positive bait box/ night (Table 16).Carcasses of rodents were discovered with symptoms consistent with anticoagulant poisoning.Three rodent species viz.Mus musculus, Rattus rattus and Rattus norvegicus were retrieved while dead from three poultry farms.Four dead rodents were found on the 2 nd night of exposure, ten after the 5 th night of exposure and increasing numbers up to nine nights (Table 16).After nine nights of exposure, there was a slight decline in intake of poison bait and the number of dead rodents.The pre-treatment and post-treatment activity monitoring showed a significant reduction (t = 8.5, df = 12, P = 0.004) in the rodent activity posttreatment period.There was 92% overall rodent activity reduction (Table 17) due to consumption of the 0.005% brodifacoum bait.

Discussion
Millet was the preferred rodent grain on poultry farms and was consumed significantly higher than wheat and rice in no-choice and multiple-choice feeding tests.The choice of food by rodents depends upon its palatability [27] and particle size [6].The ndings of this study are consistent with the previous studies conducted by [13] in poultry farms of the Rawalpindi/ Islamabad area, Pakistan.They found a strong preference for commensal (M.musculus, R. rattus and R. norvegicus) and field rodents (N.indica, M. meltada and B. bengalensis) for whole millet grains and wheat (cracked in 50: 50 w/w) with 12% eggshell powder.This study also agrees with [19], they found a high preference for Rattus rattus for millet grains in poultry farms in Jodhpur, India.M. musculus [28] and Rattus norvegicus caught from poultry sheds preferred sweet rice, proso millet, peanuts, barley and sunower seeds to corn [7].
Wheat was the second most preferred grain, and the preference for the cracked form may reect an increase in taste and chewability [29].Cracked wheat was previously noted as highly palatable to Rattus rattus on dairy and poultry farms [15,20,21].On poultry farms in India, Rattus rattus preferred cracked wheat over wheat our, poultry feed and sh meal [30].In the same way, [5] reported M. domesticus preferred soft wheat varieties (Teal, Olympic and Egret) in Australia.Cracked wheat and broken rice were also an effective bait base for control of Hystrix indica, Bandicota bengalensis, Millardia meltada and Mus spp control in the crop elds of Pakistan [31][32][33].
The use of additives can boost the affectivity of locally available food baits to become more eatable [13,14,23,31].The paired-choice test results found that the intake of bait with additives was higher than plain bait.The order of choice for baits supplemented with additives was as follows; whole, egg, bait > egg, shell, powder, bait > yeast, bait, > peanut, bait.In the order of choice among all four checked additives, baits applied with whole egg and eggshell powder stood first and second respectively.The bait use introduced with whole egg and egg shell powder was significantly higher than the plain bait in both paired choice and multiple-choice feeding experiments.In the order of choice among all four checked additives, baits applied with whole egg and egg shell powder stood first and second respectively.In our studies, the choice of whole eggs over other additives can be attributed to their flavour and calcium components, which can have a beneficial impact on bone metabolism [34].An analysis in Karachi on poultry rat food preferences (R. rattus) showed that feed grains containing 49% wheat flour and 49% broken rice were appreciably favoured with egg yolk (2%) and yeast (2%) as flavour additives over plain bait [35].Karachi (which deals with beef, fish, vegetables, dry and fresh fruits, poultry and groceries) another study on Rattus norvegicus captured from the Empress market showed first preference for yeast (2%) added bait and second preference for egg shell (2%) added bait [36,37] tested yeast powder, egg yolk and egg shell against commensal (R. norvegicus) and field rodent (B.bengalensis); egg shell powder against field rodents viz.B. bengalensis, M. meltada, N. indica, and Mus spp by [38]; peanut butter, fish mea1 and egg shell powder against B. bengalensis by [23], egg (whole) and yeast powder against commensal (R. rattus) and field rodents (N.iindica and B. bengalensis) by [14,32,39].Selection of proper bait base is vital to the efficacy of a rodenticide [17].A grain bait formulation is required to be tested for its potential to facilitate the intake of lethal quantities of poison in the pest species.For testing the field efficacy of present bait formulation consisting of millet and broken wheat grains in 1:1 proportion with 5% whole egg as an additive, an acute (zinc phosphide) and anticoagulants (coumatetralyl and brodifacoum) rodenticide were selected and their success has been discussed below: The rodents were conditioned for six days to plain bait by pre-baiting before introduction to zinc phosphide bait in current trails.It is well established that through pre-baiting high mortality of rodents can be achieved with high intake of zincphosphide [40].After offering zinc phosphide bait we recorded a sharp decrease in daily mean consumption of the poison 1bait, presumably linked with bait shynesscaused 1 by a bitter taste and garlic like the smell of the toxicant [41] as rodents link the symptoms of to the bait material ingested.Application of 2% zinc phosphide bait in new trails resulted in a 56% decrease in rodent activity at poultry farms.A number of studies performed in poultry farms using zinc phoshide baits have recorded 50% control of R. rattus population in Andhra Pardash & Ludhiana (Punjab, India) [42].Whereas [15] recorded 42.7% decrease of R. rattus with 2% zinc phosphide combined with cracked wheat and 2% egg shell, powder & egg albumin.[43] Observed 62.7% reduction in activity of B. bengalensis, N. iindica and T. indica in groundnut fields in the Pothwar zone, Pakistan, using zinc phosphide (2%) wax cake and broken rice baits.The palatability of coumatetralyl was higher than zinc phosphide in this study.This could be attributed to its slow mode of action and requiring a longer period of time between the ingestion of a lethal dose and the onset of symptoms [44].Application of this poison bait (0.0375% coumatetralyl) resulted in a decline of 82% in rodent activity.Our results are highly consistent with the study of [13] where 82% control of R. rattus, R. norvegicus and M. musculus at poultry farms was recorded.In India, [45] documented a 50 to70% control of B. bengalensis in wheat and rice cops by offering raccumin mixed cracked wheat, powdered sugar and peanut oil (in 96:2:2 ratio) bait inside the burrows.[46] Achieved 100% mortality of R. rattus by feeding coumatetralyl bait (comprising of wheat flour and 2% sugar) for 10 days in no-choice test in captivity.However, the same bait and test conditions produced 100% mortality in B. bengalensis by single day exposure.While, in Pakistan the use of coumatetralyl (0.0375%) wax blocks achieved 91.4% reduction in activity of rice field rats i.e.B. bengalensis and N. iindica [47,48].Documented 84.8% and 92.2% reduction in live burrow activity of rats with burrow baiting of coumatetralyl (0.0375%) offered in broken rice with egg shell powder in rice and wheat crops, respectively.The development of second generation anticoagulant rrodenticides, namely difenacoum, brodifacoum, bromadiolone, flocoumafen and difethiolone, has improved rodent control.Compared to first-generation anticoagulant rodenticides, the second-generation are more toxic and effective and thus are used at low doses at 0.005% concentration in the bait, and 0.0025% in case of difethiolone, and are generally effective after a single dose or day's ingestion and thus require a shorter feeding period and less bait.Also they are generally effective against rodents resistant to first-generation anticoagulant rodenticides [44].In our current trial 0.005% brodifacoum poisoned bait resulted in 92% reduction in poultry rodents activity.Similar rodent control trials conducted by PARC [49] against pest population, brodifacoum (0.005%) along with additive (egg 3%) showed remarkable reduction in pest population.Brodifacoum mixed bait emerged as highly effective formulation showing 92.63% reduction in pest population followed by bromadiolone bait (89.02%) and zinc phosphide bait (83.54%).Studies using R. rrattus, B. bengalensis and T. indica have revealed that 80-100% mortality occurs after ingestion of 0.005% brodifacoum within 2-3days of exposure to the poison baits [50,51].

Conclusions
The poultry industry suffers enormous losses globally caused by rodents.Therefore, we need to improve rodent control methods such as poison baiting.The rodents are becoming more resistant to previously available poison baits and have learned to avoid them.We need to develop more effective and economical baits.Here, we experimented with different poison baits to develop effective poison baits.The whole millet and cracked wheat with egg were found to be more effective bases for delivering the lethal quantities of anticoagulant Table 17.Data on 1pre-treatment1 and 1 post-treatment1 tracking1 tiles reflecting1decrease in rodent1 activity 1 as a result1 of 0.0051% brodifacoum m mixed1 in bait1 base 1 (millet1-wheat1 1 1:10 with 15% whole egg) at three1 selected1 poultry1 farms of Haripur, Pakistan.

Days1 of observation1
Tracking1 tiles1 Decrease in rodent activity evaluated following [26] https://doi.org/10.1371/journal.pone.0272397.t017 brodifacoum.We suggest that there is a need for continuous improvement of the poison baits over time.

Table 2 . Analysis of variance of bait takes showing variation due to grain type, replicate/farm and days of feeding.
https://doi.org/10.1371/journal.pone.0272397.t002