|Table of Contents|
Table of Contents
Introduction and design of experiments
Methods of spraying
Examination of fruit
Methods of recording data
Methods of determining effect of tests on trees and fruit
Spray deposits and presentation of data
Figures 1 through 6
STATE PLANT BOARD
UN! TS S ~TT DPA2RIMNT OF AGE!C1LMM Agricultural Research Admninistratioa Bureau of Ratonology and Plant Qurantine
EXPERIMUL MEIROD6 IN XALING ORCHRD TESTS FORE CODlING MOTH OONTX)L IN mRE WEST
Bly 3. J. imonoer and F. P. Deane Division of Fruit Insect; Invostigatione, and C. C. Q(moisi,
Division of Insectioide Investigations&!/
Introducotion *....oe....,.,.*.. 2
Deuip of experiment *...*..*.*e*** 2
Tarlationi within & tree .. 2
Treo-to-tree variation .......*..** 3
Number sad arrangenent of repe,ates ........, 4
Methods Of sPraying 6...............
~Imntion of fruit *............. 7
Preliminary examinations .e...*.e*. 7
2dimed frilt *eeeoee.e.ooeoe 8
Dropped fruit *.........e.... 9
Harvest fruit *e*......~eo~~e 9
Accuracy of observers ........... 11
Methods of rzioorfling data ............. 13
Field roorc sheet .............. 14
Bla~ry ghost .es* e*seee 1*e5~000
Methods of determinint effect of tests
on trees and fruit *............... 16
Spra deposits .eo*e~e*oee~~e~ee 17
Presekittion Of dAfts *e..e.@seeoeee 17
aumry ***..*....*.**.**ee*. 18
Litmraitro cited *...........ees. 20
~ /Aeknowledgsionts are made to Be A* Porter,
go R, Van Leeswen, F. M. Valy sad Ml. A. Ythers, of the
Burma of ltonology and Plant Quarantine. and to K. We
Bbcock. ftomrly of the Bureau, for *W helpful suggestions
and for assistance in secuiring data.
IN TEDUC TION
The codling moth (Carpoapsa pomonella L.) is the
most in curious insect eneam of apples and pears in the United States. Hyslop (86) estimates that the average anual lqss caused by this insect in the United States is $13,500,000, and that the average annual cost of controlling it is $17,500,000, making a total cost of $31,000,000. Mueh expert. mental work, designed to improve the means of control, is therefore constantly being done. It is thus important that aoourate end dependable experimental methods be used by investigators.
A combination laboratory and field method for testing codling moth inseotioides has be~e in use for some years by Steiner (8) with very dependable results. Several methods adaptable for use in the orchard alone have been employed, notably those described by Marshall and Groves (7), Cutright and Diets (2), and Hanasberry and Richardson (4).
The purpose of this eiroular is to present the methods used by the Bureau of atomology and Plant Quarantine in orobards in the Pacific Northwest, in the belief that they will be of value to other investigators in this field, particularly in the West.
DESIGN OF EPERIMTS
Variation is an inherent part of biological asperiments. here is also the variation which the investigator introduooe, and the effeoot of which he wishes to measure, In order to do so, he anst know something about the inherent variation la order to reduce it as mnoh as possible.
In experimenting with the codling moth, the inherent 'variation duoe to the host, or tree, a often be greatly reduoed by choosing trooees for th eporiment that are as near alike as possible in variety, aso, vigor, and load of trait. Even when this has been done, there is still a 'vriation in infestation, both from tree to tree and within the tree.
Variation within a tree
Variation within the tree refers to the different
degrees of infestation that my be found in different parts of the tree. rm~it in the tops of sprayed trees is often wormieor than that in the lower parts. For ample, Childs (1) indicated in 1920 that apples gnn abo-ve the 12-tot level on sprayed trees my be from 3to ies as woa as apples
rowing less than 12 feet from the vroa
Ap.les treai a miber of Dow trees in three sprayed
plate were smnsed at Yakia, Wash., separate records being kept of the fruit above and belew a lUne 10 feet tres the greu.d The tres were 20 to 2 teet high. The maber of worms per 100 apples in the upper parts of the trees ranged trem 1.9 to 7.3 times as maWay as in the lower parts, probably beeanse the upper parts were less thoroughly sprayed. There was very little difference however, in the total muber of injuries (wors and stinge) in the upper and leer parts of the trees, the upper parts having an average of let times as any ntajuries per 100 apples as the lower parts. This weas, of course, because there were fever stings Ia tfait over 10 feet from the ground than in that in the lowever parts of the tree. This Ilndieatee strongly that nost of the differene in infestatiozn between the horisontal halves of the trees weas due to the difference I spraying rather than to amy inherent difference. There is some evidence, also, that there ny be eesoe differmo e in inafestation in the different vertical partions of a tree. hesoe variations make it necessary, in making determinations of the infestation resulting from different spray treatments, to use a representative lot of fruit from all parts of the tree.
Tree-to-tree variation is illustrated in figure 1, A, iieh represents am apple orchard, the individual trees being represented by the percentages of wor frait on those trees at harvest. The trees were sprayed as uniformly as possible with lead arsenate, and yet there is not only much variation in the percentage of uora fruit between the center of the orchard and the left and right edges of it, but also some varlation between adjacent trees.
Most of these trees were sprayed in this mnner for two
successive seasons in order to find out ifether certain trees have a tendency to be worw year after year. The percentages of wormy frait on O0 of these trees in the two years had a positive oorrelatiem of 0.40. Even though this is significant, the excepetions to this teedney comprised at least 50 percent of the trees (fig. 2,A), and this would interfere seriously with trying to determine any esapeoted infestation of a particular tree.
It might be thought that this lack of high correlation
was due to variation in the elsie of the crop, and som of it my have boee, as there was much more variation of this kind the seowand year than the first. The coefficient of correlation between worminess and eaise of orop the first year wue only -0.145, uhieh is not sigifieat, and the second year it ws -0.9412, whioh is soanii eant. Ite first year the orop on 80 percent of the trees
raged betaeen 1.500 and 2#700 apples per tree, not a very great vrati on, but Ini the second, year it ranged betwean 2W0 and 1,600 apples per tree an the an- percentage of trees, a smeh greater variation* Afparently not an& eerrelation betreeenworminess and crop aiso way be expected, lesss the variation in th. latter Is considerably sore then 100 pereent,, as It wa the sooond year.
A comparison was a" of the worsiness on the trees
having a variation In crop nine of not over 100 pereent In the to succeeding years (fig* 2,B), and it Is evident that-tk. degree of worine.. on these lEres the second year eeuld net have been accurately anticipated*
Variation between trees, Is greater them variation within trees, and it is therefore important to hays, the trees usod for masuring the value of a treatment scattered at random. throughout the atperimental orchard, ad to use on& trees to ever~eose most of this variations
Numaber and Arrangement of Rapliostes
A study was made of the waber and arag et of Ieplcat". in the uniformly sprayed orchard just diseussed6 Pbur arrangement of seven hypothetical breamants were made (fig. 1. A Is C, ant D) to determine wicha m voild result In the least &theet variations. Sizee all the tree* were sprayed alike, these hypothetical treatments were not aetually different, anm therefore, If they were properly arranged,, there should be s signifi cant difference between them, In %Aite of the great tree. to-tree variation in this orchard. TWolve trees were Insluded In each of the seve hypothetical treabmnts in arnMaaginto A, X and C, thus using al11 of the 84 tres, but -In! Deoly S trees,chsn at random, were inaluded In Z&e treatment. In aetua practice it is often desirable to omit certain trees vhisk my be mo1l or of some other variety. or wiia my have a. light cro, se this arrangement Is net an unuuial one. Sie tree. sot used in D are indicated by the letter X.
An analysis of variance was then made for oah arrangeSmt, since this stat~qtioal method separaites the variation due to different causes. Y1 Arrangemnt A Is an areplisated i*.tup of sov treatments, sek composed a? a eompeat group of 2 troesI arrangemn B is similar except that the 1U trees. ame for
1/ r further infoxnation en the analysis of waiftsee and on correlation the reader is referred to 18tatistieal. btWmdslo by Goo* W. Snedecor, published by Collegiate Press, Ina*,, Am a, Iow.
the most part in a double row. Statistical analyses show a highly signpifiamt difference between treatments in both arrangement (table 1), which cannot be true, since all trees were sprayed alike. This treatment difference shows that there is an area differenoe in wrm population within the orchard. Hance neither of these arrangements can be depended on to show the actual value of treatments, although they have been such used in the past and are still used to one extent. The need of replication is evident.
One of the better replicated arrangements is that shown in C, where each of seven treatments comprises four replicates of tree trees each. No significant difference can be shown between the treatments in this arrangement; and the difference between the four blooks, each of which contains one replicate of each treatment, only approaches significance.
Table 1.--nalysis of variance of four
arrangements of codling moth treatments in an apple orchard.
Degrees Sumn of Mean
A eSource of variation Deof ees of Mean
ment freedom squares square F
A Ibtal 85 3,486 --
Between treatments a 1,289 215 7.41**
Between trees within
treatments 77 2,197 29
B Tbtal 83 5,486 --
Between treatments 6 1,523 221 7.89**
Between trees within
treatments 77 2,163 28
C Ibtal 83 3,486
Between treatments 6 550 92 o1.88
Between books 5 452 151 53.08
Interaction (error) 18 873 49 -M
Within replicates 56 1,611 29
D Total 55 1,917 --
Between blooks 7 595 85 3.042**
Between treatments 6 280 47 1.88
Interaction (error) 42 1,043 25
**Bignificant at odds of 99 to 1.
Rplanationt 1he mean square is the variance, and F is the ratio of varianoe due to two causes In A for example, 215 is the varianoe between treatments nd 29 is the variance due to other causes; 215 divided by 29 gives 7.41, vhich is highly significant at odds of 99 to 1.
The greater variance between parts of an orchard than between adjacent trees, mentioned earlier, end often noted in stadios of sodling-moth control, has led to the use of a still different set-up, shown in D (fig. 2). Hero each treatment is replicated eight times on sngle trees, scattered over the orchard so that one tree is in eaoh of the eight block* ahoun in the figure. From the standpoint of aocuraoy, this arrangement is fully as good as C, if not better; and it is a better arrangement from a practical standpoint, since only two-thirds as many trees are required to get dependable results. It has been successfully used for two seasons, in making comparisons of 7 to 12 treatments. It is probable that a sanoewhat larger number of treatments could be oopared accurately, although the number should be limited to that can be studied adequately. Unsprayed *heoks are usually unnecessary and are undesirable, as they would be likely to become so wormy that they would Influence the results. A standard treatment, suh as one comnly used by growers in the area, should be included as a check on the experimental treatmnts.
By using an arrangement such as D, then, accurate results may be obtained in an orchard having much variation in infestation, ranging in this case from 1 percent to 39 percent wormy fruit.
METHODS OF SPRAYING
A portable spray machine is the most convenient type for experimental spraying in the orchard, since relatively small quantities of material may be used in it. Some accurate method of measuring less than full tanks should be provided, either in the form of a measuring stick or of petooeks set in the end of the tank at different levels. The accessory equipment and the pressure should approximate those used by pod growers in the
Under the arrangement of replicates described above, the individual trees to be sprayed alike will be scattered over the orchard* The most convenient method of 4praying such a set-up of trees is to place the spray machine at a plaee where mter is available# lay pipes from it into the experimental block, and use it as a stationary outfit. The pipes and outlets should be placed so that all trees can be reached Ath 100 or 126 feet of hose for each gun. If there are two opeators, there should be two leads of pipe so that one man can wok frm each lead.
he tine required to do this experimental spraying ill vary with the sise and number of trees used, and the equipment available. Uhen using a 300-gallon stationary outfit as described above& and spraying 8 largo trees re-qiring 30 to 40 gallons per tree, two men can complete a trea-ent in about am hour. A set-up of 7 to 12 treatments san thus be sprayed in I to 1 days.
Since different materials are used for the various
treatnts, care must be taken that the trees are sprayed with the correct material. Tags of different colors, carrying the members assigned to the treatments, are helpful in locating the trees. If materials having dissimilar appearance or odor are used in succession, the operator can easily tell vhen the new material has reached the gun. In moving the hose front an outlet near the spray tank to one farther away, it should be remebered that the material in the pipe between the two outlets is not the now material, and the more distant outlet should Iherefore be opened and the old material run out before the hose In attached*.
EX NATION OF FRUIT
Preliminary examinations of the fruit on the trees may be made at any time to determine tentatively the relative value of the treatlmitso It will suffice to walk around the tree, examining only fruits that can be reached from the ground, jotting doe at intervals in a notebook the number examined and the number that are wormy end stung. In 2 examinations of this type, made In June, records were compared from 3,200 apples examined from the ground and from 6,400 apples from the entire tree. The apples examined from the ground averaged 2o3 and 3.0 percent injured, and those from the entire tree averaged 3.5 and 4.4 percent Injured, rospectivelyo The records of the fruit examined from the ground tell the same story as those from the entire tree, seept that the percentages are lower.
An exmintion of 100 apples per tree in this manner will be sufficient. To test this, three lots of 100 apples on each of 12 trees sprayed with lead arsenate were eamined in June, with the following results s
Lot 1 00 3.2
Lot 2 0.1 2.9
Lot 3 0.1 3.8
Average 0.1 33
1he variation is not great enough to make it necessary to examine more than 100 fruits, and if less than that number can be reached from the ground, it is possible to get a very good idea of the relative infestation by examining 5 fruits from each of 8 or more trees. In these preliminary aminations, it is not possible to distinguish accurately between worms end stings, and it is usually advisable merely to record the percentage of injured fruit or the number of injuries per 100 fruit. If more acarate records are desired, the fruits may be tagged, and at subsequent examinations aoourate dterminations may be made of the injuries recorded earlier.
The thinned fruit, which is taken from the trees in Juno or July, may be disregarded in making records of infestation.
It is ordinarily removed before there is very nmuch infestation. In one instanoe, for example, the thinings from two treatments were only 0.3 and 0.6 percent woray, respectively, and yet the harvested fruits from the sane treatments were 22 and 63 persent wormy.
The quantity of fruit that is removed by thinning may range from none to at least 65 percent of the total crop, sad this variation would influence the results if records of thinning were included. In figure 3 the percentages of wormy fruit from some actual treatments are shown. In orchard A, from 28 to 40 percent of the crop was thinned from the various treatments, and there was a very high correlation between the percentage of harvested fruit and the percentage of total fruit that was wormy. In orchard B, where the range in the percentage of the erop that was thinned was greater, the correlation was not so good. Treatments 7, 9, and 10 are out of line as regards r crds from total fruit. The percentage of the erop thinned in these treatments was so high (49 to 66 percent) that it materially lowered the average infestation, as oeapared with the other treatments where the thinned fruit was only 20 to 40 percent of the total.
In orchard A there would have been no need to take records of thinnings, as the conolusione drawn would be the same whether they were taken or not, end in orchard B the large variation in the percentage of the erop thinned introduced a definite error which might have influenced the conclusions drawn, if the records from the thinnings had been included.
Separate eainations of dropped fruit, coming fro" the trees between the tine of thinning and harvesting, are usually not necessary in arid regions. The fruit does not rot rapidly and it is usually in recognisable condition at harvest, when it may be pthered, along with fruit dropped by the piekors, and put in boxes# and a proportionate quantity of it way be included in the sample taken from the crop as a whole.
Hansberry (3). examining data from 48 apple trees in
low, found a very 'high correlation between vorminess with and without the inclusion of drops,* and concluded that dependable results oould be obtained by ea inin only the harvested fruit* At Yakima, Wash., an examination of the separate records of dropped and harvested apples from 41 trees showed a correlation between worminoss, with and without the inclusion of the dropped fruit, of 0.949, which is highly signifioat and thus substantiates the above conclusion. Separate examination of the dropped fruit therefore appears to be =necessary, except that it might be advisable if there is a large amount of it or if som other unusual condition exists.
Harvested fruit should be examined rapidly in order that the grower may remve it promptly from the orchard* It is important that only enough be examined to get an accurate determination of the infostation or other effect of the treatment, as handling is likely to affect the quality somewhat. Sinoo variation within the tree is loss than that between trees, it is less important to examine all the fruit on each tree than to examine some fruit from all the trees, provided a representative sample is taken.
In the West it is customary to pick fruit into a bag or bucket holding about a bushel, which is then emptied into a bushel box. Thus each box will normally oontail fruit from a restricted portion of the tree, and a sample made up of about equal mnmbors from eaoh box will be representative of the entire tree* If the dropped fruit is placed in boxes before the Semple i8 taken, this part of the crop is also included. Since the actual number of fruits per box will vary somewhat, it would be a little more accurate to take an equal volume from each box. If done carefully, this ould be slower than taking an equal number., and the latter method has boon found to be accurate by oMparing it with the method of examining all the fruits from a good may trees.
An estimate of the total number of fruits on the tree should be obtained before the sample is chosen, since oomparisons of the size of the crop in the different treatments are desirable. Counts were made of various proportions of the total crop on sprayed trees compared with the actual total, and it was found that if the fruit in 25 percent of the boxes, taken at random, was counted, the error would not average over 40 fruits per thousand. Since the error might be either plus or minus, the average error for a number of trees receiving the same treatment should not be over 1 percent, which is not large enough to affect the results materially. It it therefore suffietent to count 25 percent of the fruit to obtain an accurate estimate of crop size.
After this count has been made, the sample may be taken,
*Ad there is an advantage in having each sample consist of exactly the same number of fruits, as computation of the data is simplifi#1. Any reasonable number will do, and the Bureau of Matonology and Plant Quarantine has used samples of 250 fruits for determining the percentage wormy and stung, and samples of 50 injured fruits for determining the number of worms and stings per hundred fruits. If a more accurate estimate of the latter is desired, all the injured fruits in the sample of 250 may be scored for number of worms and stings.
Before deciding on these numbers, a study of data from
various apple orchards was made in the Bureau, and by calculating standard deviation the sampling error of one or more lots of 50 apples taken from different population levels was determined. It was found that little practical reduction of the sampling error was obtained at any population level by using more than 250 apples to determine percentage of injuries, or more than 50 inJured apples to determine the average number of injuries per hundred apples. This conclusion was checked in the field by comparing results obtained from 250-apple samples with those obtained from the entire crop from a good many trees, as mentioned above. The sampling method was found to be accurate, workable, and economical of time.
If there are 25 boxes of fruit to be sampled, the sample of 250 apples will of course be obtained by taking 10 fruits at random from each box. If the number of boxes is not evenly divisible into 250, it is still practicable to take a sample of emctly 250 fruits. For example, if there are 14 boxes, 18 fruits are taken from each box, making 252, and then 2 are put back. If there are 19 boxes, 13 fruits are taken from each box, making 247, and then 3 more are chosen at random from the boxes. In other words, a number vhioh will most nearly total 250 is taken from each box, and then enough fruits are added or taken away to bring the sample to just 250. By working out these numbers in advance
for any number of boxes likely to be encountered and providing each observer with a card showing the numbers to be taken as well as the number to be added or subtracted, samples of 250 fruits can be taken accurately and without loss of time. A portion of such a table is shown below:
Number Number of Number Number Number of Number
or fruits taken added or of fruits taken added or
boxes from each box subtracted boxes from each box subtracted
is 14 M2 21 12 42
19 13 +3 22 1 +
20 12 +10 23 11 -3
1he saplo can be conveniently hooked by a crew of
three observers, two making the etion of the fruit end the third recording the results of the examination on tally sheets or field record sheets (fig. 4). The fruit is examined one at a time, the observer calling out the condition of each fruit so that it can be recorded. Detailed records of each injured fruit are kept on the right-hand portion of the tally sheet until 50 such fruits have been recorded. We rest are then recorded on the left-hand portion. (Details of recording the data are given on page 13 under "Methods of Recording Data"). If the fruit is lightly infested, tres observers can work about as fast as three, one of them not only examining fruit but also making the records for both. Other methods of scoring or tallying the fruit, such as by means of a battery of tally rosters, one for each type of record desired, may be used, and some of these methods are described by Cutright and Diets (2) and by Hansberry and Richardson (4).
wee men an sort out and marine a sample of 250
fruits in about 15 minutes, and the estimation of the total =umber of fruits should not take more than another 5 minutes. It will take 15 or 20 minutes for two or three men to grade the fruit for color, if that is to be done, as mentioned on page 17.
Aocuracy of Observers
fto aourecy of the observers should be carefully
chocked before the actual eamination of fruit begins, and in the case of inexperienced men some training will be neoessary. For this purpose, a lot of 100 apples, including a easiderable mber injured by the codling moth, should be laid out in rows on a table out of doors or in a room uhore there is plenty of sunlight. EAoh apple should be carefully
scored by at least two experienced observers, in the manner that is to be used in practice, and an agreement reached as to th exact number of w rsa and stings on each apple. Then the other observers should score the apples, and their resalts should be compared with the actual condition of the fruit. If eare is taken to keep the apples iu the same relative positions on the table, it is then possible to go back ad show each observer just where he made mistakes and why.
In testing the accuracy of observers, it has been
found that errors made by experienced men often cancel each other out, and that the final result on 100 or more apples is likely to be quite coourate. This is shown in table 2.
able 2.-Average percentage variation from
oorrecot score of percentages of wormy and stung apples and numbers of worns, stings, and total injuries found by six observers
Observer Worm Stung Worms Stings Injuries
A +0.4 -4.5 -2.35 -18.3 -11.3
B +0.8 -6.7 -1.5 -16.6 -10.0
C +0.0 +4.5 -7.1 -*33 -4.4
D +1.7 -5.3 +2.3 -15.5 -7.7
Average +0.7 -3.0 -2*.1 -13 5. -8.4
9 -2.9 +.7 -19.7 -6.2 -12.1
F +7.4 -14.0 +6.4 -26*.4 -12.0
Observers A to D had had several years' expeOrieneoo, and their scores did not vary from the correct score enough to affect oonolusions drawn from the results. Observer wBuas les sooaurate in sooring worms but scored the stings about as aoeurately as the more experienced men. Observer F was as accurate as B in total injuries, but had a tendenoy to call some of the stings worms, that is, he recoorded more worms and fewer stings than the others did. It was found that most of the errors made were due to moertainty as to whether to resord a given injury as a worm or a sting, or due to overlooking *all stings. Such tendencies can be more or less correoobted with a little practice. The distinction botawcen worms and stings should be thoroughly understood, as well as the importan oe of recording all stings, no matter how small. A good uay of differentiating worms from stings is to consider as worme all inJuds having a surfaee diameter of more than .oneeighth inoh or that are so freshly mad. that it is evident
the worm, in still alive, and to consider all other injuries as stings., It will help the observer to out open some apples after he has soared them in order to verify his decisions, although a practice cannot be made of doing this in the course of regular ewziations of fruit. Actually, however, 'with the scores shown in tho tables and with fruit that was 10 percent wormy, observer B would have recorded 9.7 percent wormy fruit, end observer F. 10.8 percent wormy fruit,, neither of which is a very serious error.
A very interesting point is shown in table 2, namly, that there is more accuracy in recording percentages of wormy and stung fruit than numbers of worms end stings. The experienoed observers averaged three times as far off on number of vorms as on percentage wormy, and more then four time as far off on number of stings as on percentage stung. It is also evident that the average of the records of several observers is likely to be more nearly correct than the record of any one observer. If the four experienced observers had examined about equal proportions of the sample fruit from a treatment that had 10 percent of wormy apples and 30 percent of stung apples, they would have recorded the fruit as 10.1 percent wormy end 29.1 percent stung. If 10 worms and 80 stings per hundred apples had been present, they would have recorded 9.8 worms and 26 stings.
=RhOD$ OF RE00RDI1NG DATA
The amount of data recorded from the fruit depends an the information desired. The simplest record would consist only of the total number of fruits and the total number of wormy fruits. from which the percentage of the latter could be figured.
In most tests of insecticides, however, it is importat to record both wormy and stung fruits, as some materials reduce the number of stings much more than others. This my be done by recording as wormy all wormy fruits, and as stung all fruits that are stung but not vormy,, the total of the two being the total injured fruits. But it is more accurate to record separately all the wo ray fruits end all the stung fruits, regardless of whether the other type of injury ocurs on them, and to record also the fruits that are both wosMy and stung in order that the aggregate of injured fruits my be computed. It is often desired to know, also, how many worms or stings there are per 100 fruits, and this requires a record of the total number of worm entrances and stings.
Whether the investigator should record the percentage
injured or the number of injuries per 100, or both, depends on the use he is to make of his results. If he desires them
chiefly for practical use in showing fruit gowr what to expect from various treatment, then percentages should be used, since they show the effect on the orop from the stand. point of yield. As already indicated, also, it is possible for observers to record the percentage of wormy and stoug fruits more accurately than the number of worms and stings. If the figures are to be used for king careful scientific deduotions, then the number per 100 fruits is more desirable. In the case of heavy infestations, the intensity of attack can be better differentiated, and it in also possible to calculate from those numbers the total number of worms end stings per tree, should that information be desired.
As a matter of faot, the correlation between percentage of wormy fruits and umber ofworms per 100 fruits$ eto. is so high that usually the some oonolusions. are drawn from either method. Pbr example, the records from 219 trees havlng a maxim of 25 percent of wormy apples showed the highly significant correlation between perountage wormy and number of worms per 100 apples of 0.950.
Field Record Sheet
Printed or mimeographed field record sheets are convenient for recording data in the orchard, especially if tally registers are not used, as mentioned abovo. A form commonly used in the Bureau of Bntomology and Plant Quarantine is shown in figure 4, and this has been filled out in fa*ile of an actual record shoot from an insetioide experiment. A supply of theso shoots can be carried on a olip board, a separate shot being used fbr eaoh tro.
At the top of the shoot are the Initials of the observer iho made the record, followed by those of an observer who assisted* This record is sometimes of use in correcting errors or doubtful entries. The other entries at the top are self-explaatory, "plat" in this case being synonomous with 'treatment.W
Men the e tion of a sample is begun, each injured
fruit is scored on the basis of actual umber of worse (W) and stings (8), and those ar recorded in the right-hand block of squares until 50 injured fruits have been examined. fach vertical pair of squares is used for recording 1 fruit. In this case, the first 3 injured apples examined each had 1 worm, the fourth had 2 worms and 1 sting eto. The remIning fruits are then merely scored as "wrmy'n tI), 'stung' (8) or.-'both' (B),8 and these records are placed in the left-hand block of squares, In this ase, miong tho apples that were left, there were 6 wormy, 2 stung, and 2 both morny and stung.
Previously, the number of boxes and number of fruits counted have been recorded below the right-hand block of squares In this instance there were 20 boxes of picked apples and 3 boxes of windfalls, the latter totalling 657 apples; and 5 boxes containing 903 apples wore counted. These figures. and the number of apples in the sample, are
all th;Lt need be entered during the field exwmnaion.
At some convenient time, usually after the harvest is over, the computations shown are made. First, the total worms and total stings in the right-hand block are calculated, as well as the total wormy and stung fruits and the averages, expressed as worms per wormy apple, stings per stung apple, and injuries per injured apple. Then the total wormy and stung, in this case 36 and 24, are entered in the appropriate aolums in the left-hand block, and the total clean, wormy, and stung apples in the entire sample are calculated, as well as
the percentages. In this example there were 44 wormy and 28 stung, or a total of 72; but two apples in the left-band block and 10 apples in the right-hand block, or a total of 12, were recorded as both wormy and stung, hence there were only 60 different wormy and stung apples, leaving 190 clean.
In the lower part of the form the average number of apples per bushel is entered, computed from picked fruit only, since dropped fruit is usually shriveled and small, and also the total numbers of apples in the various categDries are shown. "Total worms," for example, is obtained by mltiplying 'apples wormy" by "average worms per wormy apple," in this case 751 x 1.33 3 999. A simpler form could be used if only percentages were desire or if tally registers were used.
Further calculations are made on another form, called a sumnary sheet (fig. 5). This sheet may be used for summarizing data from dropped fruit and harvest examinations from individual trees, and also for bringing together the data from all trees treated alike. In the example, the data from 8 trees sprayed with a treatment that was numbered 9 are shown. Referring to tree No. 8 it will be seen that the figUres, with the exception Of those under the heading "Average No. per 100 apples," have been copied from the field record sheet (fig. 4). These averages are obtained by dividing the total number of apples into the total of worms, etc. In this case the equation is 999 + 4269 x 100 23,4 worms per 100 apples The treatment totals are then calculated, the percentages being simple averages. The heading "Average No. per injured apple" refers to the type of injury listed in the subheading, and means "worms per wormy apple"# "stings per stung
0-1, XTV. PLANTs~ wOI
-W 16 40
apple, and *Injuries per injured apple". If, however, thiis shoot is being used to obtain the season totals of samples of dropped and harvested fruit from a single tree, these percentages and averages must be weighted on the basis of the o~mbr of fruits in each samples in order to give each samrpling Its proper importance. The average yield in number of apples and number of bushels, and the site of the fruit, shown at the bottom of the sheet, are all simple averages, the latter being the average of the number of apples per bushel shomi on the field record sheets for the trees included in the summary.
The method of obtaining data of this type is not accurate enough to warrant showing percentages or "average No, per 100 apples' to more than one decimal place. 'The "average No. per injured apple" should be calculated to two decimal places, however, in order to increase the accuracy of the calculations made from these figures. Much time may be saved by using a calculating machine for the various computations.o All results should be checked by another individual, and products and quotients may be quite accurately checked with a slide rule.
NEIROlDS OF DETEMANING EFFECT OF TESTS ON TREES AND FRUIT
Now materials or combinations should first be tested on a tree or two, or even on a branch if there is a possibility of more than nominal injury from them. The tests can them be enlarged if major injury does not occur. Materials should be tested on all the oivn varieties grown in the region, since Instances have 'been recorded of a material causing practically no injury to some varieties of apples and serious injury to, others. Observations should be made for several seasons and in more than one orchard, because effects of the materials are very likely to be influenced by weather, cultural conditions, and tue vigor of the trees.
Frequent examinations should be made in order not to overlook any dropping of a portion of the foliage that may occur, which might result in smaller fruit. Pranature dropping of the fruit should also be observed.
Some materials may affect the size or color of the fruit, and a rough Idea of the relative size may be obtained by *&alQu3stiag the average number ef fruits per picked box from the counts that have been made. A more accurate method of comparing not only the size but also the rate of growth consists of making periodical measurements of' selected fruits at intervals from the time thinning is done until harvest* This is done by taking the circumference with a steel tape graduated to millimeters. Harley and )Aasure (5) and others have used the rate of fruit enlargement 'as an index of tree response to various factor* affecting tree growth and behavior," a~nd they consider it suffi-
clent to have complete records from 25 normal fruits from each treatment, distributed about equally on several trqes, and as nearly the same size as possible 'when chosen, To insure having this number at the end of the season, 30 to 40 fruits should be Included to start with. Figure 6 shows the results of such a series of measurements from two treatmentsr and it indicates that the effect on size was a gradual one in this inatance.
1he color of fruits, particularly of red apples, may be affected by some insecticides. To determine this, the harvested fruit may be graded according to grading rules in force in the region. It in advisable to use all the fruit from as many trees as possible, as trees sprayed alike often prodvjes crops with widely varying degrees of color,
In making tests of insecticides, It is very important to have a chemist working along with the entomologist* Methods of controlling the codling moth have reached the point where compatibility of materials and compcusition of emulsions and stickers are all-important* An experienced chemist in required, not only to devise effective mixtures but also to determine the deposit of the toxic material on the fruit, for on the quantity of this deposit is dependent to a considerable degree the control obtained. For this purpose, chemical analy, ses should be made before and after each application of spray:!
The development of spreaders and stickers may be cited as an example of the value of such chemical analyses. In this way it was learned that some spreaders caused the toxic materials to run off the fruit, thus defeating their purpose* Means of correcting this fault were worked out, and now, for example, the deposit and control obtained from 1 pound of phenothiazine in 100 gallons are just as good as were obtained d a few years ago from 3 pounds.
PRESENTATION OF DATA
Sufficient general information about the tests should be given to enable the reader to judge for himself the value of the results* This should include number and dates of applioations, methods of mixing sprays and of spraying, method of tak-. iug results, and any pertinent information about the weather or other factors that might influence the results. Any effect on the trees or crop should be mentioned. The materials used
It is not within the scope of this circular to discuss the methods of mcalyses or of collecting samples for analyses.
should be suffioiently described so the reader will know just what they are. Only results that are discussed in the text should be shown in tables. A compact way of presenting a season' s results from three orchards is shown in table 3 (page 19).
This table shows, among other things, that in two of the three orohards there was no significant difference in results from lead arsenate with spreader and tank-mix nicotine bentonite, as regards number of worms per 100 apples, but that there were fewer total injuries in the latter treatment, and in orchard C the number of these was significantly less. The conclusion from these particular tests muld be that nicotine bentonite, as used, was as good as lead arsenate with spreader in preventing vorminess of fruit, and somewhat better in preventing stings. Percentages of wormy and injured apples could be shown in the same manner, but it should not be noeessary to use both methods of presenting results in the same publication.
Because of the large amount of experimental work being dose on the eodling moth, it is desirable to have somewhat standardized methods of making orchard tests of insecticides. Depndable comparisons my be made by spraying single tree plots replicated eight times, and a convenient way of doing this is to lay pipes in the experifaental block and use a portable spray machine as a stationary outfit. Preliminary indications of the value of the test materials may be obtained by exsmining 100 fruits per tree without removing then. Thinned fruit need not be examined, but fruit dropping later should be looked at unless it will remain in recognizable condition at harvest. Final determination of infestations may be made by examining a representative sample of 250 fruits from each tree. Two experienced men can make this examination in 15 or 20 minutes. It is desirable also to make an estimate of the total number of fruits per tree and of any damage the insecticide may have done to the fruit on the tree. Data may be recorded on field record sheets and the desired computations made later and transferred to summary sheets for permanent record. Chemical analyses of spray deposits should also be made. Data may be presented in capact tabular form by showing only the nuer of worms and injuries per 100 fruits, and also showing the differences required for signifioance at odds of 19 to 1.
14 Sala ID H
M vi r4
t*00-% or4 a 0 $3
"4 0 FA
M a 0
p 94 0 r4
.p sit r.4
.4 d o so -P 0%-0 0
a r4 0 40
11 "4 so 4
d 0 d
S- 020LI TERATURE CI TED
(1) Childs, L.
1920. Spray gun versus rod and dust in apple orchard
pest control. Oregon Agrioultural Uzpt. Sta.
Bul. 171, 46 pp., illus.
(2) Outright, C. R., and Diets, H. F.
1933. The technique of oodling moth field experiments.
Jour. loon. Ent. 26: 392-401, ilues.
(8) Banbsherry, T. R.
1936. An investigation of codling moth populations as
they affect control experiments. Thesis submitted to the graduate faculty of Iowa State
College for the degree of Dootor of Philosophy.
(Unpublished.) Abstract in Iowa State College
Jour. Sci. 11s 68-66.
(4) Hanaberry, T. R.. and Richardson, C. H.
1935. A design for testing technique in codling moth
spray experiments. Iowa State College Jour.
Sci. 10t 27-55.
(5) Harley, C. P., and Masure, M. P.
1938. Relation of atmospherio conditions to enlargement
rate and periodicity of Winesap apples. Jour.
Agr. Research 57: 109-1238, illus.
(6) Hyslop, J. A.
1938. Losses occasioned by insects, mites, and ticks
in the United States. U. S. Dept. Agr., Bur.
Ent, and Plant Qnar. -444, 57 pp. (Processed.)
(7) Marshall, J., and Groves, K.
1936, Field methods for investigation of codling moth
insecticides. Jour. Boon. Ent. 291 1137-1144, illus.
(8) Steiner, L. F.
1939. The laboratory-rield method for testing codling
moth insecticides. U. S. Dept. Agr., Bur. Ent.
and Plant Quar. L-488. 10 pp., illus. (Prooessed.)
00 03 P1 40-=~ I
9, *~b0 0 9@ K 0K
----- ..... x a x x 43
0 1--- ---in in x. K -9 9 94 P4
x~ x9 4 00K
. . .. .. I ....K. .K
0 x x*
...... ..... ....
a~~ '-, ox
0 x xe~h
_._. _._.. ._ ..... 4 4
949x *A0 90 00
049 0 QP 4>II
20 = Q1 r4 00
91 t1 910110
On__ 4 on_ ___in_
40 30 20 to 0 10 20 30 40
figure 2.-A, Percentages of wormy apples on 80 trees uniforuly sprayed with lead arsenate in two sooessive years; B, same,
including only trees having variation in crop alse of not over
0 0 r4
0 0 0 0 0
co 0 In CY
Figure 4-- COMING WNT IMIEBTIOATIONS
Field Record Sheet
Naae or observer A(4 D-C C ate OCT 2 2 1942
Typ. of eaminatioa HARVEST Locality YakimxA, Waat. orur Le Vesconte
Variety __ __ __ __ __ __ __ __ Plut Reliation ro
Total and Total and
Number of vorq -eren -Number of wreand Ave""g
-_and atm c~ls-_lean wom st131 ~rInjured ape
--~~~~~ap --- //012 I f~
By 2N 2 21 /
8 __ 2_ 2) 1 /1 1 6
W 1 1 .2 7
B W13 31 1 1/10
------------ aTotal inj uries -Z
B --- --- -Total injured 3( 2Z5(e
Total ~ ?Average 1133 /2Z/
Percent 70/ //2 Averige injuries S/
-eos nj" 1Z10-Per injured apple
No. Apples SMo. Bushels
in sape, Z6-6 in lot eapl ed_2. 67)
per bse %.( ane
Apple* ocne 74)3
Apples Total '7Apples: Clean J '
Wormw / Total WOrMB 5
Stung 7 Total Stngs 7
Injure Z.~ 5-F
1"m 2S 2 g 16 93 30
JJJZLW -,W SEPT.
9 mwIddW gpm of an'* .s prmy6