AN APPROACH TO ALFALFA WEEVII. MANAGEMENT IN MICHIGAN Thesis for the Degree of M. S. MICHIGAN STATE UNIVERSITY RICHARD A. CASAGRANDE 1971 ......... III!IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII L I I N University AN APPROACH TO ALFAIFA WEEVIL MANAGEMENT IN MICHIGAN By: Richard A)“Casagrande A THESIS. Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE [Bpartment of Entomology 1971 ACKNOWIEDGEMENTS I am more than appreciative for the support and assist- ance that Dr. F. N. Stehr has provided me in my program. Dr. D. In Haynes has participated in many conversations which have been of great help in the design and analysis of my reu search. The degree of cooperation extended to me by these two individuals has been more than one could ask for. Other members of my guidance committee have also been very helpful. Dr. G. E. Guyer has provided a source of ' continual Optimism and encourazement. Dr. R. F. Ruppel, who has been following the alfalfa weevil since it reached Michi« gan has been very helpful to me as has Dr. M. B. Tesar of the CrOpand Soil Science Department. .Both have provided me with new vieWpoints on the problem. I I Dr. G. G. Gyrisco of the Department of Entomology at Cornell University, who has had much experience working on the alfalfa weevil, was of great help in the early develOpe ment of my research. A source of many interesting conversa- tions and much encouragement has been Dr. C. D. F. Miller of the Harrow Research laboratory at Harrow, Ontario. Stuart Gage and Bill Ruesink have helped me in some form just about every day and I would certainly like to ac- knowledge their assistance. i1 Finally I must thank Harold Webster of the Kellogg Gull lake laboratories Experimental Farm who COOperated with me to the fullest extent in the imnlementation of my program at Gull lake. TABlE OF CONTENTS IIST or TABLES...........................................€%§e IIST OF FIGURE............................................vi INTRODUCTION...............................................l DESCRIPTION OF FIELDS STUDIED..............................3 LIEE HISTORY IN MICHIGAN...................................5 overWinterj-ng Eggs.00......OOOOOOOOOOOCOOOOO0,0.0IOOO‘ll ParaSites in MiChiga-n.OOOOOOOOOOOOOOOOOOOOOOO0.0.0.0016 METHQDS 0F COllECTING-AND ANALYZING DATA..................18 sampling for EggSOOOOOOOOOOOOOO...’09‘OQOO‘OOOOOGOIOOJ-{3 Mpling for larvanOOOOOOOOOOOOOO0.09.000000000000002]. Sampling for Pupae...................................25 Sampling for Adults..................................23 Freeisj.on O.f DataOOOOOOOOOOOOOOOOCOOO0.0.0.009I00000025 DevelOpmental Times of Eggs, larvae, Pupae and Parasites.............................28 Methods of Analysis..................................54 AGE SPECTLFIC MORTAHTYOOOOOOOOOOOOOOOOOGO0.0.00.00.06.0000u6 DeUSj-ty Ibpendant MortalityOOOo.OOOOOOOOOOOOOOOOOOOOOIA'6 Effect of Cutting on weevils.........................47 Effect Of cutting on ParaSiteSOOOOOO0.00.00.00.000000’49 PMSNAGD‘IG AHnAEAOOOOOOOOOOOOOOOOOOOOOOOQ0.0000......0.0.0.51 Summary of Management Practices......................56 HERAWRE CIWOOOOOODOOOOOOOOO0.0.0....OOOOOOOOOOOOtOOOOO58 LIST OF TABLES Table Page 1 Overwintering Eggs (1969-1970)...,.,,,,,,,,,,,,,...,.1N 2 Precision of Sample Data for 1971 (Standard Errors Expressed as Percentages of Means at Peak Densities).26 3 Ibveloplrfiental Times.00....OOOOOOOOOOOOOOOOOOOOO0.0.0.29 4 Results of 1970......................................35 5 Results of 1971......................................36 113 OF FIGURES F1 gure ’ Page 1 DenSity of adults in the alfalfa field StudiedatGUIlIflkeill1970;........................8 2 Density of eggs and larvae combined throughout the 1970 season...........................9 3 Density of eggs and larvae combined throungout t1183.97l 388.801]...coooocooooooooooooooooolo A Relationship between stem density and egg density...20 5 variance of samples for immature weevils............27 6 Standard errors that result from the use of different thresholds in computing degree day requ1136rnents from Table .§OOCCOOOGOOOOOOOOOCOOO0.00003]- 7 Density of eggs in field A throughout thelg7l seasorlo000.000.000.00...003.00.000'00000000037_ 8 Density of larvae in field A throughout the1971seasonIOOOOOOOOAOOOOOOOOO...0......OOOOOOOOIQO 9 Density of pupae in field A throughout the 1971 season.....................................44 10 Density of parasitized larvae in field A throughout the 1971 season..........................45 vi INTRODUCHION The alfalfa weevil, Hyperagpostica (Gyllenhal) (Coleop- tera, Curculionidae), has been a serious pest of alfalfa in the United States since l904 when it was first introduced near salt lake City, Utah (Titus, 1907). It Spread through- out the west but was never found east of the Great Plains A until 1950 when it was discovered in Maryland (Bissell 1952). The east coast pOpulation spread much more rapidly than the western one, and this difference, in addition to several others noted by Blickenstaff (1965 , Koehler and Gyrisco (1961), and Armbrust et. al., (1970) indicates the two strains were possibly introduced from different areas of Europe (where it is native). The first report of the alfalfa weevil in Michigan was in 1966 (Dowdy 1966). Since that time it has spread rapidly throughout the state and by 1971 was present in potentially damaging numbers throughout the entire lower peninsula. The alfalfa weevil is the most serious pest of alfalfa in Michigan, primarily because of larval feeding damage. larvae skeletonize the leaves and reportedly have the poten- tial to destroy a stand if large numbers are present after a field is cut by continually eating back the new growth. To date alfalfa weevil damage in Michigan has been reduced primarily by insecticide use. Of Michigan's 1.25 million acres of alfalfa, 206,000 acres were treated with pesticides for weevil control in 1970 (Anonymous, 1971) at a cost of 8 - 12 dollars per acre for materials, equipment, and labor (Janes and Ruppel, 1969), not including an unknown cost to the environment associated with the production, distribu- tion, and side effects of these pesticides. My overall objective has been to develop an effective management program for the alfalfa weevil in Michigan which would minimize insecticide use and provide for satisfactory alfalfa production. To this end I have spent three field seasons studying the weevil. The first (1969) was devoted primarily to learning about the weevil, the crop, and the parasites and developing objectives and sampling techniques for further research. The 1970 season provided sound infor» mation on the life history of weevils and parasites with rela- tion to cropping practices and allowed for the development of a preliminary management scheme which was implemented and evaluated in 1971. It is not the intent of this thesis to provide a chronoé logical record of these three years of research. ”Instead I am presenting only those results which are pertinent to the development, implementation, and evaluation of the management program in the hOpe that this might provide the basis for a , state-wide program and suggest means for evaluation and sub- sequent refinement of such a pregram. DESCRIPTION OF FIELDS STUDIED In 1970 two fields were studied, one at Gull lake, and one at Collins Road. Both were mature stands of Vernal alfalfa and neither had been treated with insecticides for at least one year. The field at Collins Road was a fairly clean, dense stand of alfalfa, situated on level ground. A 100 X 200 foot research plot was located near its center. The Gull lake field was of similar density with slightly less weeds and was situated on a slight south 310pe, although ' the research plot selected within this field (100 X 200 foot) was on relatively level ground. On 26 May one half of the Gull Lake plot was out, leaving the remainder of the field uncut, and on 28 May the Collins Road plot was similarly out. Neither field showed any degree of larval feeding damage at cutting date. In 1971 all research was conducted at Gull lake. Three fields were selected for close study. Field A, the same field studied in 1970, was a somewhat sparse, four year old, stand of vernal alfalfa about three acres in size.‘ Field B was a two year old stand of Saranac alfalfa, somewhat denser than field A and 2.5 acres in size. ield C was a three acre stand of one year old Saranac alfalfa that was of moderate density and vertually free of weeds, the latter feature being in marked contrast with the other fields. Field C was sprayed on 30 April with malathion to kill ovi- ‘positing adult weevils, thus establishing a low density of eggs. On 28 May 25% of field B was out in the form of a strip (fifteen feet wide) extending for the length of the field. V 0n 3 June another such strip was cut in field B and fields A and C were cut as well. In the center of field C a strip eight feet wide was left uncut, extending from one end of the field to the other. In field A a similar strip was left in the center (measuring three feet wide) and a three foot strip ' was left uncut along each of two sides of the field as well. Fields B and C showed a very slight degree of larval feeding damage at the time of cutting, but field A was considerably damaged. IIFE HISTORY IN MICHIGAN. Obviously the first step in the develOpment of a pest management program is a complete knowledge of the life his~ tory of the pest. The life history of the alfalfa weevil was reasonably well understood in 1969 as outlined in Extension Bulletin E~639 (Janes and Ruppel 1969). This was based on some limited field observations and on projections based on data collected in neighboring states (particularly Ohio). Since the basic bi010gy of the weevil was known, I was able. to concentrate on those specific aspects influenced by Michi~ gan's northern climate. . Alfalfa weevils are active in the very early spring and have been observed in flight even on warm winter days (ProkOpy and Gyrisco 1965). I have not observed any such activity in Michigan before mid-April, however considerable evidence of a much earlier period of flight activity has been observed for three consecutive years. During the first week in April of 1969, large numbers of alfalfa weevils were observed on the sand dunes along the lake Michigan shoreline in Berrien ’ County (D. L. Haynes, pers. comm.). On 8 April and 27 April, 1970 I made similar observations at the same location and on 1 April, 1971, I again found large numbers of weevils on the sand at Grand Haven, Michigan. A week later I returned to \I‘I Grand Haven, finding weevils there and at several other loca~ tions as far south as Holland, Michigan. These weevils were apparently flying on the warmer days of late March and early April and were concentrated at the shoreline by the cold air over the water. In 1970 and 1971 several of these weevils were collected and returned to the laboratory fer observation. It was found in both years that, although they fed readily on greenhouse alfalfa immediately after their capture, no oviposition occur- red until two weeks after the initiation of feeding. weevils which were given water but no food did not oviposit at all despite the availability of suitable oviposition sites. Since no alfalfa was yet available in the field when these weevils were collected, it was concluded that a feeding period was required before oviposition could occur. This was found to be in agreement with Snow's observa- tions (1928) that sexual maturation of dispensing overwinter- ing alfalfa weevils does not occur until they feed after breaking diapause. Samples taken in research plots at the EntomOIOgy Re- search Station on Collins Road in East Lansing, Michigan (hereafter referred to as Collins Road) and at the Kellogg Gull lake laboratories Experimental Farm in northeastern Kalamazoo County, Michigan (hereafter referred to as Gull lake) in early April, 1970, showed that most alfalfa weevils did not overwinter in the alfalfa fields. According to Hamlin et. al., (1949) and Manglitz (1958) many alfalfa weevils overw star in woods, hedgerows, and field borders. These overwintering weevils apparently fly about on warmer days and, after the first new growth occurs in the alfalfa fields, begin to concentrate in these fields. As Shown in Figure l, weevils were first found in the Gull lake alfalfa field in 1970 on 17 April (the first sample date after the‘ new alfalfa growth began). Although the weevils were present in the field on 17 April, no significant oviposition was ob- served until 30 April despite suitable oviposition sites in dead stems from the time of their arrival. Similarly in 1971, although the new green alfalfa was available as early as 8 April, no oviposition was observed until 20 April. This 'seems to indicate that the weevils in alfalfa fields, like the weevils from lake Michigan, require a feeding period be- fore oviposition begins. Once oviposition begins, the egg density increases quite rapidly and the peak egg density occurs about the third week of May (Figures 2 and 3) after which the rate of hatching exceeds that of oviposition and egg density subsequently de- clines to nearly zero by late JUne. As these eggs hatch, the first instar larvae begin feeding on the alfalfa buds but ~ cause little noticeable damage to the crop. The first visible damage occurs just before the time of the peak larval density 313“] 02/9 08/9 OI/9 02/9 02/9 2 i7 *9 s8 40'! «8| 4f?! Ol/‘r 9/17 .[ 1' ' 05/9 09/17 03/17 ,,,a' e Figure 1. --Density of adults in the alfalfa field studied at Gull Lake in 1970. whdo \m 9}, on} cm} 03» owe. OE ome owe 9e on}, 8 TI}. a _ _ _ _ 9 q 100 .. 3 9 nu _ -oo. 3 av 1 N 0 .109 w I no Odom m2} _Ou a it A .63.. jam " ll. . N mkdo 0275.30 u u jOONW vlqwa 4<>m<4 u 4 Z x$ : g 3 353 .Q‘g o 4 D a 2 EEE \V' uJ.Jocnu..u.u_ unlau I' O '. m‘.” uJ .J L) (n . ' <§~ I . .11; L 1 n n 41.1 \— O O O O 0 ¢ (9 CD C) (D C) ID V‘ r0 N "" "Bu/awaw ow sees Figure 3. --Density of eggs and larvae combined throughout the 1971 seaspn. 11 (about the first to second week of June at these sites) when .the earliest larvae reach the third and fourth instars and begin eating a great deal. larvae pass through four instars in about two weeks and pupation lasts about five days. After emergence adults feed for a few days and leave the field, not to return until the following Spring. It should be noted that the ates given for peak egg and larval densities apply generally to the lansing and Gull lake areas and would vary considerably to the north and south. In addition there can be a great deal of variation between even adjacent fields, depending primarily on the density of the stand and the 810pe of the field. At Gull lake in 1971, the larval pepulations in one of the fields with a sparse stand of alfalfa on a southwest facing lepe reached its peak twelve days before a level field with a denser stand. Furthermore a field with a steeper southern slope and sparser stand had an even earlier larval pOpulation, preceeding the former by approximately one week. Over-wintering Eggs 0f considerable interest to a management program is the question of overwintering eggs. In Delaware adult weevils return to alfalfa fields in the fall, where after feeding for a while they lay considerable numbers of eggs, most of which hatch early the following spring, causing early damage 12 to the alfalfa (Burbutus et. al., 1967). Similar observa- .tions have been made in Virginia (Neodside et. al., 1968), New Jersey (Dively 1970) and New York (Armbrust et. al., 1966), but at the more northerly sites, generally fewer eggs were noted to survive the winter. Ruppel and Janes (1969) predicted that only a few overwintering eggs were expected to survive Michigan's winters. But Armbrust (1966) found that in New Yerk overwintering survival varied from 8.8% to 91.6%, depending on a number of factors including snow cover and field location. Townsend and Yendol (1968) noted consid- erable differences in survival between eggs laid in upright -and in lodged alfalfa stems in Pennsylvania. Thus there was considerable evidence that micro-climate 'as important in determining overwintering survival. To determine how fall-laid eggs.survived Michigan's winters, a fairly comprehensive study was initiated in Octo- ber, 1969. Ten samples of all the stems in a square foot of alfalfa were taken at the-Gull lake and Collins Road fields twice each month. All stems were split open and the eggs were counted. Also recorded was the viability of eggs, the depth of snow cover on each sample date, whether stems with egg masses were lodged or erect, the height of egg masses in erect stems, and the diameter of stems with egg masses. In addition, percent parasitism of eggs was noted. These obser~ vations, which were continued through January, 1970, showed 13 .one important fact. In Michigan, alfalfa weevils simply do not oviposit in the fall. The results presented in Table 1 indicate that very few eggs were laid in the fall and of these few, only a small number remained viable through the winter. Meanwhile, Niemczyk, (1970) published the results of his observations in Ohio. He found considerable fall oviposition and overwintering survival in southern Ohio, but in the north fewer eggs were laid in the fall, and of those laid, relatively few survived the winter. Subsequent samples I took in mid-April, 1971 showed a continuation of this trend into Michigan. On the Ohio-Michigan border in lenawee County there was a mean density of 7.9 over- wintering eggs/square foot in four alfalfa fields while the mean density at Gull lake and Collins Road was 2.2 and l.u eggs/square foot respectively. The data I collected, when combined with that presented by Niemczyk, indicates a trend toward less fall oviposition and less winter survival as one moves north. It does not seem surprising that very few eggs survive the cold of the winter, but the reason for decreasing fall oviposition with increasing northerly latitude is less apparent. A review of the literature suggested several possible explanations but the key factor seemed to be adult diapause. Guerra and Bishop (1962) showed that alfalfa weevil females 1% mancmpowpn nmsawe¢ * \ e3 ROOH o.o. : _ o.o nae em 0.0 0.0 nn< ea 0.0. 0.0 nee 0H 0.0 due new mm 0.0 0.0 see m 0.0 0.0 one we 0.0 0.0 one m 0.0 0.0 eoz ea 0.0 0.0 see e 0.0 o.H see me 0.0 0.0 nee m &0 H.n 0.m a: . . m.a _ . e.m an 0.0 m.m mo 0.0 0.0 w 0.0 0.0 Rem. . o.m . .e.m . 0.0 0.0 mam m.H . m.m - o.o m.~ acne o.o e.o eem o.o e.a nae mm nn4 0H nae s new ea ash co é >oz ea >02 H #00 mH 900 H anaaannee mam m u0\mmpfimmnnm mpM\nwmm open Sign; mam emufinefinennm mnfinwwm 38 neom neaaaoo mme Hana Aoemfiumomav nmwm.meaneenaeneeo .H mHmsa l5 ,undergo a true ovarian diapause. Huggans and Blickenstaff (195s) noted that field pepulations of weevils consisted of both dispensing and nen~diapausing adults and felt that this trait could be selected for. The fact that very few eggs are laid in the fall in Michigan can thus he explained by the selective pressure for diapausing adults eXerted by the severe winter. Those few adults which do not enter diapause feed in the early fall and oviposit, but most of these eggs do not survive the winter and the tendency for fall oviposi- tion is selected out. This explanation also seems adequate to explain the gradient in fall oviposition indicated by Niemczyk (1970). In the southern part of Ohio where eggs can survive the winter there is much fall oviposition, but farther north relatively fewer weevil eggs survive the winter, resulting in a shift of the population toward spring oviposi— tion. The majority of weevils in Michigan leave the alfalfa fields shortly after emergence and enter diapause, not to emerge until the next spring. These are comparable to the "ditch bank" weevils which Snow (1928) observed to remain sexually immature until spring. The length of diapause of those weevils which do undergo diapause was found by Huggins and Blickenstaff (1964) to be determined by the photOperiod eXposure in the larval stage. They noted, as did Rosenthal and Koehler (1968), that longer daylength exposure in larvae results in longer diapauses by 16 the adults. There is a one hour difference in daylength between southern Ohio and central Michigan_during the peak (larval periods on mid-May (Ohio) and mid-June (Michigan) (list, 1951) so it is probable that photoperiod determines diapause and hence time of oviposition. Additional data on photoperiod responses would be interesting, but at present the best estimate of the duration of diapause in Michigan is the 170 days (at twelve hours larval photoperiod) noted by Huggins (1964). A 170 day diapause would be more than ade- quate to prevent fall oviposition in Michigan, as weevils could not become active until January. After breaking diapause in January these weevils remain sexually immature until after they initiate feeding in April. Parasites in Michigan Several parasites have been introduced into Michigan in recent years. Many of these have been recovered and are thought to be established, however by 1971 only Bathyplectys curculionis (Thompson) (Ichneumonidae) was present in densi- ties high enough to have any effect on weevil pepulations. In May of 1970 and early June of 1971 Microctonus aethiops (Nees) (Braconidae) was recovered at Gull lake following releases made in 1969 (Stehr and Casagrande 1971). Bathy- ‘plectys anurus (Thompson) (Ichneumonidae) and Tetrastichus incertus Ratzburg (Eulophidae) were recovered at Collins Road -4 17 in 1970 following releases in 1969. In addition Tetrastichus incertus was also recovered at Gull lake in 1971 where it was Ireleased in 1968. Two other parasites, which were not introduced into Michigan, were discovered in the research plots. These are both egg parasites and at present do not appear to be of any real significance. In sampling for overwintering eggs many Anaphesgpraetensis_(Foerster) were recovered. These are re- corded in Table 1 along with the egg density. Since over- wintering eggs are not significant in this area of Michigan these parasites probably will not be of much importance, ex: cept to add to the already severe selective pressure exerted by the Michigan winter. The other egg parasite recovered, Patasson lune (Girault), was recovered in fair numbers from eggs collected in Ju1y. This was found to exert up to 60% parasitism on 2 July, 1970 but egg densities are very low at that time. 'This parasite is described by Brunson and Coles (1968) as primarily a parasite of overwintering eggs and is apparently not very important in Michigan. METHODS OF COllECTING AND ANAIXZING DATA In order to intelligently interfere with the life cycle of the alfalfa weevil in the form of a management program,. it is necessary to first understand the natural population regulators. Once these factors are understood it is possible to try various manipulations that might reduce weevil popula- ‘ tions. Thus it was necessary for me at an early date to de» velOp sampling techniques that would allow for an understandu ing of the sources of natural control. Subsequently, for evaluating management programs, a long term pOpulation dynamics study seems to be the best approach and it is for this purpose that I am presenting a discussion on sampling techniques so that it may provide a ready reference for future research. Sampling for Eggs Other researchers have sampled for alfalfa weevil eggs by examining all the stems in a unit area or by sampling a certain number of stems selected at random throughout a field. Results have thus been eXpressed as numbers of eggs per square . foot or as numbers per stem. Both of these sample units seem useful, however they serve different purposes. Numbers of eggs per stem might be adequate for predicting larval damage, I and indeed, expressing larval density in terms of stem density 18 19 might lead to a hatter understanding of larval competition. However, for a pepulation dynamics study it is advantageous to express the density of all stages in the same units and it does not seem logical to express pupal or adult density as numbers per stem since they are more often found on the ground than on_the plants. Samples for all stages were thus taken on a unit area basis and stem density was later recorded. Since many authors, including Armbrust et. al., (1960), Blickenstaff (1966), and Woodside et. al., (1968) expressed egg density in terms of stems, I found it desirable to de- termine the relationship between stem and egg density in order to interpret their work and to decide on the prOper sample unit. I took 34 square foot samples on 18 June, 1970 at Collins Road and recorded the number of eggs and the number of stems in each. The results (Figure 4) indicate that for the normal range of stem densities, the number of ' eggs is independent of the number of stems in a sample. Thus eggs are distributed by area and not by stems and it is most reasonable to sample according to the same distribution. In choosing the prOper size area for each sample, I decided to make each sample representative of the alfalfa field as a whole. Since each plant is usually several inches from other plants I found a square foot to be the minimum size sample which provided an adequate representation of spatial distributiOn of the plants. Furthermore, consideration of 20 1 O —1 O 0 o O 0 0 . O O O. . 0 0 g . -+ o. : O .. .. a O + 0 .4 O l J _l l l l l c) (D C) C) C) C) C) q- <_\g Q 00 «0 - 'd" m 331/8993 Figure 4. --Re1ationship between stem density and egg density. HQ 50 STEMS/fiz .the time and effort required to process square foot samples in light of available resources snowed such samples to be practi- cal from an economic standpoint. Thus one square foot was decided upon as the basic sample unit. Eggs were sampled by removing all stems of all types from the area delimited by a square foot frame placed in a randomly selected location within a plot. living stems were clipped at ground level and dead stems were also collected and placed in a plastic bag. Stems were later split with a single-edge razor blade and examined for eggs. Data recorded included the numbers of eggs, egg masses, and stems in each sample. Frequent samples of eggs were tested for viability. These were held on moist filter paper in small plastic petri dishes and kept at 75°F. Eggs were treated with .01% captan to retard fungal growth. ' Samplinggfor larvae ' Different methods were used for larval sampling in 1970 and 1971. In 1970 I took one half square yard samples, by placing a sample frame over the foliage, clipping it, and placing it in a plastic bag, taking care not to shake the larvae out. I then placed. the samples in large Berlese funnels, each equipped with a 500 watt lamp and collected the larvae in a jar of 70% alcohol for 24 hours. This method of sampling is fast, requires relatively little labor, and P0 I‘D _provides for consistent results. However, it has a severe shortcoming in that considerably less than 100% of the larvae in the field are counted in a sample. It is not possible to clip and remove the intertwined alfalfa stems from the field without shaking or brushing some larvae from the plants. ’Careful examination of the ground on two dates revealed that no first or second instars were knocked from the foliage, but 18% of the third and 25% of the fourth instar larvae were lost in this manner. A further reduction occurred with the Berlese funnels from which the percent recovery'.of the first, second, third, and fourth instar larvae was 27, 90, 70, and 70-respectively. Thus in order to determine the actual den- sity in the field, the results obtained had to be corrected by two factors. This seemed to be a very undesirable means of sampling so in 1971 samples were designed to eliminate these losses and the necessary correction factors. In 1971 a threeesided (U—shaped) one foot2 sample frame was constructed of 1/8" x 3/ " flat iron, with a piece of fine white fabric measuring 2h" x 24" attached to the back. In taking samples, this frame was slipped between the stems at ground level and the entire square foot of foliage was gently leaned over the cloth, clipped, and placed in a plastic bag. Any larvae which were dislodged from the plants by clipping . were easily found on the material and added to the sample. In addition the soil surface and alfalfa crowns within the .sample frame were carefully examined and any larvae found were added to the sample. The samples were then returned to the laboratory where each stem was carefully examined and the larvae were removed and preserved in 75% ethyl alcohol. Before discarding the sample, it was individually washed in 95% alcohol to remove any larvae missed. The few larvae which were found were counted and added to the other larvae from the sample. This method of sampling requires considerably more effort than the method used in 1970, but has the advantage of allowing complete recovery of all the larvae from the field. Sampling for Pupae In 1971, the same samples were taken for larvae and pupae. After all the foliage was clipped and placed in the sample bag, the crowns of the plants and the litter were carefully examined for pupae and any found were added to the sample. Samples were then taken back to the laboratory where they were frozen and held until they could be processed by hand and the pupae and larvae counted. Sampling_for Adults Anyone who doubts the effectiveness of evolution in pro- viding for effective coloration and behavioral adaptations should try sampling for adult alfalfa weevils. For almost 24 .seventy years the problems of sampling for adult weevils have hampered researchers in this country. Probably the best tech- nique that I have used for determining absolute adult density involves the use of a square foot metal frame 4" high on each side and sharpened on the bottom edge. The frame is pressed into the soil and adults on the foliage within the frame are shaken off into it and the stems are then clipped at ground level and removed. A very sharp lawn-weeding tool is then used to cut the roots of each plant, the crowns are removed and examined for adults, and the soil within the frame is swept up with a small wisk broom, placed in a plastic bag, and taken back to the laboratory where the adults are separ- ated from the soil by floatation in water. Although this technique has not been used extensively, it seems to be quite effective, particularly if the samples are processed while the adults are still alive. . Adults were not sampled in the 1971 study on age spec- ific mortality because it was thought that pupal density would provide an adequate estimate of adult density. Since there are no pupal parasites, apparently low predation, and low pupal mortality, the large effort involved in sampling for summer adults (particularly with separating emergence from dispersal) seemed not worthwhile. 25 _Precision of Data Most researchers working on population Studies agree that the standard error of a sample should approximate 10% of the mean. This is an arbitrary figure which seems to be a reasonable compromise between accuracy of results and effort required in sampling. In the egg data I collected in 1970 I found that the magnitude of the standard error is dependant upon the magnitude of the mean. Thus to keep the-standard error at 10% of the mean it is necessary to take larger sam- ples at low densities than at high. Using the results of 1970, I determined that ten_l foot2 samples would provide ’for quite precise data at the expected peak egg densities. Taking into consideration the time and expense involved in 'taking and processing each square foot sample, the large number of fields studied and frequency of sampling, I con~ cluded that for 1971 I could not take more than ten samples per field without reducing the number of fields or sample frequency. Although this sample size seemed adequate for high densities, it was not expected to provide great pre- cision at low densities. However, a precise knowledge of density was far more important to me at higher densities then at the low and thus samples consisting of ten square feet seemed to be an adequate Compromise between precision and excessive expense. Similarly a constant sample size of ten 1 foot2 samples was chosen for larvae and pupae with the 33 U\ realization that the results would likely be less precise than the egg data because of lower densities. Figure 5 shows the means and variances of samples taken in 1971. A regression line fit to the data points on this figure shows the relationship between means and variances to be 2 z 0.3692g+ 1.458 log i log 8 The variance for a given mean can either be read directly off Figure 5 or calculated according to the equation S2 a .6234 +~antilog (1.458 log E) which incorporates the necessary conversion factor for trans; forming logarithmic data back to arithmetic as described by Bliss (1967). Table 2 shows the precision of the estimates of peak density in each field for each stage sampled during the 19. season. ThBlE 2. Precision of Sample Data for 1971 (Standard Errors Expressed as Percentages of Means at Peak Densities) Field A Field B - Field 0 Eggs 9.1% - ' 10.5% 16.4% larvae ' 11.2% 12.2% 15.4% Pupae ~ 19.2% 3.7% 27.5% 27 MEAN oeusn‘v (2) Figure 5. --Variance of samples for immature weevils. . 100,000 , , , , I0,000 ~ -i T": 1,000 _ .4 DJ 22 £3. Cf. —' -1 q IOO :> m . .J G. 3 IO _ ~+ 0’ . EGGS o LARVAE a PUPAE . 1.0 .. - 0| i l . l 0.! LC 10 IOO I,000 10,000 28 Developmental Times of 3338, larvae, Pupae and Parasites w ~ “our..- ;- - s.” C -——-—* In order to make any calculations on age mortality or to compare densities between fields or years, it is necessary to determine the total incidence for the season of whatever stage is under consideration. 1 accomplished this by using the method of T. R. E. Southwood (1966) in which the density is plotted on a graph throughout the season and the area under the curve is divided by the developmental time, giving the total incidence per unit area for the season. Developmental times of alfalfa weevil life stages are readily available in the literature in papers by Sweetman and Wedemeyer (1935), Koehler and Gyrisco (1961) and Roberts et. al., (1970). In addition H. D. I‘liemczyk kindly provided some additional unpublished egg data that.he collected. The_ combined results of all these sources (presented in Table 3) show that deve10pmental times are determined by temperature ‘ exposure. Since field temperatures change dramatically from mid-April to Ju1y (and occasionally equally as dramatically from one day to the next), I decided to calculate develop- mental times in terms of degree days, thus eliminating the variation caused by temperature changes. 'In order to calculate degree day requirements, it is first necessary to determine a lower temperature threshold, above which degree days are accumulated. This is typically done by plotting ercent deve10pment per day over different 29 TABLE 3. Developmental Times Tbmp. (00.) ' Days as Eggs Days as larvae Days as Pupae 6.9a 90.5 - 10.0a 44.5 4 12.00 49.3 80.5 12.8a 29.0 4 13.5d 29.0 ? 15.0a 19.3 9 17.0.0 19.8 28.6 18.98 13.0 - 20.00 13.4 20.5 21.16 11.0 — 22.0b 9.3 >15.2 23.7b 8.5 1 25.3b 7.2 - 26.7d 9.0- 4 27.0b 6.9 10.6 . 28.0c 6.0 9.8 28.7b 6.2 - 50.3b 5.8- - 32.0b “5.5 8.1 36.00 5.1 9.6 37.0b . 4.0 . 10.9 : §$::::i% :23 5:3:::%.£1%18%3) c Koehler and Gyrisco (1961) d fiRoberts et. al., (1970) 11.5 9.9 6.7 6.4 5.1 4.9 30 temperatures, finding the point at which the regression line crossed the X axis, and defining that point as the lower threshold (that temperature below which no development can occur). This method seems objectionable to me for two reasons. First I do not think one can justify extending a. regression line past data points and secondly, I do not think it is biologically meaningful to establish an exact threshold such as 44.50F. (7.200.) (Roberts et. al., 1970) on the asSump- tion that no development occurs below that temperature. It seems quite possible to me that different physiological processes could have different thresholds and hence it might not make sense to establish a fixed threshold. Thus I decided to use whatever threshold would provide the best results for degree day requirements for each stage. This was accomplished by arbitrarily substituting different. thresholds and calculating degree days for each different experimental temperature in Table 3. The mean number of degree days and standard error was then calculated for all- temperatures at each threshold and these standard errors were plotted against the thresholds (Figure 6). The bottom point of each curve was-associated with the threshold which gave the most consistent results and was thus selected as the lower temperature threshold for deve10pment. Thus the thresholds for eggs, larval, and pupal deve10pment were determined to be 9°C., ll.5°C., and 9°C. respectively. Using s EGGS c LARVAE A PUPAE l 0d (3-9) Asossa OHVONVLS Figure 6. --Standard errors that result from the use of different thresholds in computing degree day requirements from Table 3. I4 l2 IO (“Cl IHRESHOLD .the data again in Table 3, the mean degree day requirements above these thresholds were determined to be 150.6 for eggs, 171.6 for larvae, and 73.8 for pupae. . With this information available, it was then necessary to calculate the rate of degree day accumulation within the fields. Temperature records were kept during 1971 by using a hygrothermograph located at the ground surface within an alfalfa field. Degree day accumulation was calculated from the daily maximum and minimum temperatures by using the sine curve method described by Baskerville and Emin (1969) with a computer program written by Gordon Baskerville (unpublished). ' Calculations were made by uSing hBOF. (900.) and 52°F. (1100.) as thresholds and the rate of degree day accumulation for the entire season was determined for each threshold. Unfortun- ately such hygrothermograph records were not kept during 1970 so I looked up the daily maximum and minimum temperatures for Gull lake and East lensing in the Climatological Record (pub- lished by the U. S. Department of Commerce). Then using hygrothermograph records for JUne 1970, measured at the sur- face of a mixed field of alfalfa and oats by S. H. Gage, I deve10ped conversion factors for converting the standard temperature records to surface temperatures in an alfalfa field. Using the equations: Maximum temperature at surface 3 air temperature X .73-+ 16.34 \N \4 Minimum temperature at surface : air temperature X .65 +- 225 ' I then corrected the air temperatures from the Climatological Record to the temperatures at the surface of the field and proceeded to determine the rate of-degree day accumulation for both thresholds again using the computer program written by G. Baskerville. The effectiveness of this technique for determining deve10pmental times was checked on two occasions during the 1971 season by inserting freshly laid eggs into alfalfa stems in the field and checking them daily to determine the time -until hatch. Eggs laid on 15 May hatched on 26 hay (144 degree days later) and eggs laid on 4 June hatched on 12" June, requiring 152 degree days. These values are in close agreement with the predicted requirement of 130.6 degree days for egg hatch. Not much information was available on the deve10pmental times of larvae parasitized by Bathyplectys curculionis, how- ever Armbrust et. al., (1970) showed that they spend less time in the feeding stage than unparasitized larvae. Using his results I determined that parasitized larvae spend 84% as long in the feeding stage as unparasitized larvae. Using the larval threshold of 11.50d. I thus determined that parasitized larvae require lt4.l degree days for development as Opposed to the 171.6 required for unparasitized larvae. 34 Methods of Analysis Cutting an alfalfa field has a complex effect on alfalfa .weevils and parasites. It not only causes the removal of eggs and mortality of larvae and immature parasites as described by Hamlin et. al., (1947), but it also causes a reduction in oviposition by both weevils and parasites. In order to evalu- ate these effects for different cutting dates, I found it necessary to analyze graphs of density plotted against time for each stage in each field. The results, presented in Tables 4 and 5, are the product of this analysis which seems best explained by discussing a complete example. The total incidence in the uncut portion of field A was calculated by the method of Southwood (1966) which involved measuring the area under the outer curve in Figure 7 (130,952.62 egg degree days) and dividing by the deve10p- mental time of eggs (130.6 degree days). The result 1002.7 .eggs/foot2 is the average number of eggs.laid in each square foot of field A during the period studied. Marking the time of cutting (506 degree days) on this curve allowed the determination of density at cutting time, showing that 160.0 eggs/foot2 were exposed to cutting (16% of the total incidence). To determine how many eggs hatched before cutting I subtracted the deve10pmental time of eggs (130.6 degree days) from the time of cutting (506 degree days). Thus any eggs TfiBlE 4. Results of 1970 EGGS Hatched Before Cutting Date Exposed to Cutting Removed by Cutting laid in Cut Part After Cutting laid in Uncut Part After Cutting Total Incidence in UnCut Part~ Total Incidence in Cut Part IARVAE Pupated Before Cutting Date Exposed to Cutting Killed by Cutting Hatched in Cut Part After Cutting Hatched in Uncut After Cutting Total Incidence in Uncut Part Total Incidence in Cut Part 1 2cut on 26 May cut on 28 May 812.4 0.0 12.0 12.0 43.4 94.7 111.5 45.0 Gull lakel Collins Road2 u —-—.——— 175.3 129.5 86.1 23.0 155.0 454.0 (37') t3 .'.. ,’L o 0.0 7.5 7.5 45.0 .100.1 105.3 43.4 TABlE 5. Results of 1971 2 a Field A Field Bl Field s2 Field o“ zeccs _ Hatched Before Cutting Date 729.2 .259.6 436.0 91.1 Exposed to Cutting 160.0 195.0 114.0 4.0 Removed by Cutting 125.0 176.2 78.0 0.0 laid in Cut Part After Cutting 8g.5 74.8 59.8 5.5 laid in Uncut After Cutting 12 .1 151.3 111.9 5.4 Total Incidence in Uncut Part 1002.7 646.5 646.5 99.2 Total Incidence in Cut Part 842.2 359.9 543.6 101.0 IARVAE Pupated Before Cutting 44.0 0.0 1.0 3.0 Exposed to Cutting .21 .0 24.0 39.2 61.6 Killed by Cutting 11 .3 24.0 39.2 35.9 Hatched in Cut After Cutting 16.1 54.2 30.1 3.0 Hatched in Uncut After Cutting 38.7 154.7 120.8 2.9 Total Incidence in Uncut Part 23 .1 165.9 165.9 68.5 Total Incidence in Cut Part 7 .5 67.1 34.8 23.6 PUPAE Emerged Before Cutting 0.0 0.0 0.0 0.0 Exposed to Cutting 0.2 0.0 0.0 0.0 Total Incidence in Uncut Part 8g.8 89.2 83.2 31.4 Total Incidence in Cut Part .5 ' 5.1 .l 2.1 PARASITIZED IARVAE Pupated Before Cutting 1.0 0.0 2.1 3.0 EXposed to Cutting 15.2 6.8 2.9 9.2 Killed by Cutting 2.7 - .9 6.6 Input in Cut After Cutting 12.1 16.0 13.7 1.2 Input in Uncut After Cutting 16.1 36.4' 30.9 3.9 Total Incidence in Uncut Part 26.5 40.3 40.3 19.5 Total Incidence in Cut Part 19.9 21.7 17.5 6.6 gout on 28 May cut on 3 June 37 momwuuoqoxmmmIP ©me mom 000. com 00m , 00.... 00m 00m m>