UNIVERSITY OF CALIFORNIA PUBLICATIONS COLLEGE OF AGRICULTURE AGRICULTURAL EXPERIMENT STATION BERKELEY, CALIFORNIA THE COMMON HONEY BEE AS AN AGENT IN PRUNE POLLINATION Second Eeport BY A. H. HENDRICKSON BULLETIN No. 291 January, 1918 UNIVERSITY OF CALIFORNIA PRFSS BERKELEY 1918 Benjamin Ide Wheeler, President of the University. EXPEEIMENT STATION STAFF HEADS OF DIVISIONS Thomas Forsyth Hunt, Director. Edward J. Wickson, Horticulture (Emeritus). Herbert J. Webber, Director Citrus Experiment Station; Plant Hrocdiny Hubert E. Van Norman, Vice-Director; Dairy Management. William A. Setchell, Botany. Myer E. Jaffa, Nutrition. *Eobert H. Loughridge, Soil Chemistry and Physics (Emeritus) Charles W. Woodworth, Entomology. Ealph E. Smith, Plant Pathology. J. Eliot Coit, Citriculture. John W. Gilmore, Agronomy. Charles F. Shaw, Soil Technology. John W. Gregg, Landscape Gardening and Floriculture. Frederic T. Bioletti, Viticulture and Enology. Warren T. Clarke, Agricultural Extension. John S. Burd, Agricultural Chemistry. Charles B. Lipman, Soil Chemistry and Bacteriology. Clarence M. Haring, Veterinary Science and Bacteriology Ernest B. Babcock, Genetics. Gordon H. True, Animal Husbandry. James T. Barrett, Plant Pathology. Fritz W. Woll, Animal Nutrition. Walter Mulford, Forestry. W. P. Kelley, Agricultural Chemistry. H. J. Quayle, Entomology. Elwood Mead, Eural Institutions. J. B. Davidson, Agricultural Engineering. H. S. Eeed, Plant Physiology. D. T. Mason, Forestry. C. L. Eoadhouse, Dairy Industry. fFRANK Adams, Irrigation Investigations. W. L. Howard, Pomology. O. J. Kern, Agricultural Education. John E. Dougherty, Poultry Husbandry. S. S. Eogers, Olericulture. David N. Morgan, Assistant to the Director. Mrs. D. L. Bunnell, Librarian. Division of Pomology W. L. Howai d W. P. Tufts E. H. Taylor Eichard Schmidt W. L. Sweet Guy L. Philp A. H. Hendrickson Milo Wood * Died July 1, 1917. t In co-operation with office of Public Eoads and Eural Engineering, U. S. Department of Agriculture. THE COMMON HONEY BEE AS AN AGENT IN PRUNE POLLINATION BY A. H. HENDRICKSON INTRODUCTION A second report on the part played by the honey bee in the pollination of prunes in the Santa Clara Valley is given herewith. The first report 1 showed conclusively that bees are a necessary aid to pollination with the French and Imperial varieties of prunes. It also showed that while one or both varieties, but particularly the former, may set a fair crop when but few bees are present, they produce a heavier crop when bees are numerous in the orchard throughout the blossoming season. It further showed that prac- tically no fruit was produced when all pollen-carrying insects were kept from visiting the blossoms. At the same time it raised the question why the Imperial prune tree did not respond to pollination by bees in a manner similar to the response of the French tree. It left unsolved, however, the question as to whether the French or the Imperial prune could set a crop of fruit, when worked upon by bees which were not allowed to come into contact with any other variety. PLAN OF SECOND YEAR'S WORK The work with the honey bees was done in connection with a large pollination project being carried on by the Division of Pomology of the University of California. As in the case of the previous year, the experiment was performed on the Sorosis Ranch, situated in Santa Clara County, about three miles northeast of Saratoga on Saratoga Avenue. The orchard is made up principally of French (Agen) prunes, eighteen to twenty-five years old. Four rows of Imperials topworked on Robe de Sergeant run cross-ways of the part of the orchard situated on the north side of the road. The trees are on mixed rootstocks ; some on almond, some on peach, and others on Myrobalan. The soil is a well-drained clay loam, inclined to be some- what gravelly in places, particularly in the section occupied by the Imperials. The general condition of the trees, with the possible exception of some of the Imperials, is better than the average in the i Cal. Exp. Sta. Bui. 274. 216 UNIVERSITY OF CALIFORNIA EXPERIMENT STATION district. A cover crop of barley and vetch is plowed under every spring, and barnyard manure from the dairy, run in connection with the ranch, is applied liberally every two or three years. The trees are given a regular dormant spray and pruned lightly each year. The experiment was conducted by erecting tents of white mosquito bar over the experimental trees so that bees and other pollen-carrying insects could be excluded or confined. The experiment of 1916, in which one pair of adjoining French and Imperial prune trees were confined in a tent with a hive of bees, and one pair of trees of the Fig. 1. — Tent of white mosquito bar erected over a pair of adjoining French and Imperial prune trees. same varieties kept free from all bees, was repeated. In addition, one French tree and one Imperial tree, respectively, were encased in single tents with bees. All of the trees used were as nearly the same size and in the same condition of health and vigor as it was possible to secure them. The double tents over the pairs of adjoining trees were about 40 feet long, 25 feet wide, and 15 feet high (Fig. 1). The single tents were 16 feet long, 16 feet wide, and 15 feet high (Fig. 2). They were as nearly insect-proof as possible. Three trees each of the French and Imperial varieties were used. Two double tents were erected, each enclosing a pair of adjoining THE COMMON HONEY BEE IN PRUNE POLLINATION 217 French and Imperial trees. Single tents housed the two remaining trees, a French and an Imperial, nsed in the experiment. Thus, four tents were erected to enclose the six trees grouped in the combinations mentioned. In one tent, enclosing a pair of French and Imperial trees, a hive of bees was kept for six days during the blossoming period (Fig. 3). All pollen-carrying insects were excluded from the second double tent which enclosed the other pair of French and Imperial trees. A hive of bees was kept in each of the single tents, Fig. 2. — Tent of white mosquito bar erected over a single tree. which enclosed a French and an Imperial tree, respectively. It was thought that practically all of the questions arising from the use of bees in an orchard where these two varieties of prunes were planted would in this way be tested. Under the tent in which there were no bees, the effect of the absence of pollen-carrying agents was tested. In the tent enclosing both a French and an Imperial tree, the question of inter-fertility or the necessity for interplanting for purposes of cross-pollination between the two varieties was studied. In the single tents the ability of each variety to set fruit with its. own pollen was tried out. In addition to the experiments outlined 218 UNIVERSITY OF CALIFORNIA EXPERIMENT STATION above, counts were made of blossoms of each variety on trees growing in the open, in order to get the set of fruit that occurred under aver- age orchard conditions (Fig. 4). The last named count served as a check on the experiments. Fig. 3. — Bee hive under tent enclosing a prune tree. A cluster of bees may be seen at the entrance of the hive. The trees which had been used in the experiment during the pre- vious season were also carefully watched. In order to study the effects of the abnormal crops produced the previous season, counts of the blossoms were made to find out the amount of fruit set, and observations were also kept of the general growth of the trees. THE COMMON HONEY BEE IN PRUNE POLLINATION 219 In addition to the tests outlined, other orchards where bees were kept among the trees in 1917 were under close observation. Follow iip* i ... I £ "^ ,. *— • fl *t ■■,'::.;"■, ."-; , ■ >>!* Fig. 4. — French prune tree in full bloom, typical of kind of trees used in the experiment. ing the work of 1916, a number of prune growers in the Santa Clara Valley and in other sections of the state kept bees in their orchards during the blossoming period. A brief summary of results in each case follows : 220 UNIVERSITY OF CALIFORNIA EXPERIMENT STATION Mr. H. H. Dingley of Saratoga placed fifty hives of bees near the edge of his orchard containing about fifty acres. When this orchard was visited in August, the trees were literally breaking beneath the crop. Mr. Dingley also reports a large crop on some cherry trees nearby. Mr. J. F. Pyle of San Jose kept a number of hives of bees in his orchard. This orchard was not visited, but the owner, in a letter, said in part: "Our Jefferson plum orchard which has never borne a fair crop has a heavy crop this year . . . ; in fact, all our trees of all varieties, peaches, apples, plums and prunes, have an exceptionally heavy crop. ' ' Fig. 5. — Heavy crop on Tragedy plum tree in E. W. Skinner's orchard, Yuba City, Cal. Eesult of keeping a large number of colonies of bees in the orchard during the blossoming period. About fifty colonies of bees were kept in the orchard of Mr. C. E. Jackson of Santa Clara. When this orchard was inspected, the prune trees were loaded to capacity, and the owner was also harvesting a full crop of Washington plums from trees that had never before produced a satisfactory yield. Mr. E. W. Skinner of Yuba City placed 113 colonies of bees in his mixed orchard of about eighty acres. All varieties of fruit set exceptionally heavy crops, which was attributed by the owner to the work of bees (Fig. 5). Mr. Skin- ner with Mr. H. P. Stabler, Horticultural Commissioner of Sutter County, also conducted a number of experiments to test the work of bees in the orchard. Several branches of different kinds of fruit trees were covered with mosquito THE COMMON HONEY BEE IN PRUNE POLLINATION 221 bar so as to exclude bees and similar insects. The result was a heavy set of fruit on all parts of the several trees except the branches covered by the mosquito netting. 2 HOW THE WORK WAS DONE The mosquito-bar tents were erected over the trees chosen for the experiment before any of the flower buds had opened. The frame- work was started March 12, 1917, and the tents were completed on March 15, 1917. As the blossoms opened slowly due to the back- Fig. 6. — A few of the bee colonies in Sorosis orchard during the blossoming period. wardness of the season, it was not until March 23 that enough flowers were sufficiently advanced to justify placing the bees under the tents. The Imperials had opened three or four days ahead of the French, but there was still a large number of the blossoms of the former open when the majority of the blossoms of the latter variety reached the stage when pollination could take place. The bee colonies were placed under the tents on the evening of March 23 and removed on the evening of March 29, after having been confined for six days. During the first day or two the bees did not do much work, but occupied their 2 A full account of this test is given in the Pacific Eural Press for April 7, 1917. 222 UNIVERSITY OF CALIFORNIA- — EXPERIMENT STATION time trying to escape. By March 26, however, they had settled down, seemed contented, and kept busily at work on the blossoms. The tents were removed April 14 after all the blossoms had fallen. The weather throughout the time the bees were under the tent was ideal for pollination purposes, being warm and clear, with little or no wind. Approximately 2000 blossoms were counted on each of the several trees in the experiment. The branches on which these blossoms were situated were tagged after the method used in pollination experiments, and the number of fruits resulting from these blossoms were carefully watched throughout the season. From these figures the percentages of fruit set was obtained. A large number of blossoms were counted on trees not treated in any way to obtain the average percentage of set that prevailed in the orchard. The trees used to obtain this average orchard set were the same ones used for that purpose in 1916, so the average set for the orchard for the two years may be studied without regard to differences due to individual tree variation. RESULTS OF EXPERIMENTS The results obtained, as in 1916, were very decisive, and the figures show greater uniformity than do those obtained in the previous season's work. Tables 1 and 2 give the results of the year's work. TABLE I Behavior of French Prune Trees With and Without Cross-pollination, Season of 1917 Blossoms Matured fruits counted counted Per cent of Treatment Mar. 21, 1917 Aug. 7, 1917 fruit set French prunes: average orchard set 7272 962 13.2 French prune tree : under tent from which bees were excluded 1391 6 0.43 French prune tree: enclosed with bees and an Imperial tree 2120 330 15.5 French prune tree: enclosed with bees alone 2152 409 19.0 The average orchard set of the French prune trees used to check the work with the bees was 13.2 per cent, which is exceptionally high for the variety. This was probably due to the fact that the trees were near the Imperials and not far removed from one of the several colonies of bees placed in the orchard by the owner (Fig. 6). With the French prune tree from which all pollen-carrying insects were excluded, only 0.43 per cent of the blossoms set fruit. This result illustrates the necessity for pollen-distributing agents like bees in the orchard. THE COMMON HONEY BEE IN PRUNE POLLINATION 223 Where a French and an Imperial tree were enclosed in a tent tog-ether with a colony of bees the set of the French tree was 15.5 per cent. No doubt, practically every blossom on this tree was visited one or more times during the period the bees were in the tent. The French tree alone under a tent with bees set 19.4 per cent. This high set is also probably due to the fact that each blossom was visited a number of times. The pronounced difference between the tree which had no pollen- carrying agents at work upon it and the trees which were exposed to bees is apparent at a glance. TABLE II Behavior of Imperial Prune Tree With and Without Cross-pollination, Season of 1917 Blossoms Matured fruits counted counted Per cent of Treatment Mar. 21, 1917 Aug. 7, 1917 fruit set Imperial prunes: average orchard set 1535 112 7.2 Imperial prune tree: under tent from which bees were excluded .• 2036 7 0.34 Imperial prune tree: enclosed with bees and a French tree 2153 170 7.9 Imperial prune tree: enclosed with bees alone 2184 66 3.02 The average orchard set of the Imperial prune trees was 7.2 per cent. This percentage represented a good crop, when the profusion of blossoms on each tree is taken into consideration. The set of the Imperial tree from which all pollen-carrying insects were excluded was only 0.34 per cent. Here again, the need of a pollen-distributing agency is plainly indicated. When enclosed with a colony of bees together with a French tree, the Imperial tree set 7.9 per cent. This result shows that most of the flowers had been visited by the bees, and many of them had been caused to set fruit. The Imperial tree upon which the bees worked without coming into contact with the blossoms of any other tree showed a set of 3.02 per cent. The persistent work of the bees which visited every flower was probably responsible for this decisive percentage of set. As was the case of the French, the value of bees for pollination purposes with the Imperial is strikingly shown. The results of the work during the season of 1917 agree in general with the important results obtained in 1916. Figures 7 and 8 present a graphic comparison of the season's work. 224 UNIVERSITY OF CALIFORNIA EXPERIMENT STATION Where the French and Imperial prune trees were enclosed in the tent with bees the set of fruit was greater with the French variety than the average set for the orchard. The increase in the percentage of set was undoubtedly due to the work of the bees. During both years, the French trees from which all bees and similar insects were excluded set a negligible crop. Without insect aid, these trees were unable to set fruit. In 1917 the French tree which alone was enclosed in a tent with a French prunes /f/7 XZE2ZZZZZ&'3-*% / I /f3% 6 *** vver* * x-e** from Vv/tvc/* Sf'7 V7777Z 79K ""** **** wS/t /^ n^^i^ y ^-- ■■"■v-^^^^m^^m^^.t Fig. 9. — French prune tree showing drooping branches and scarcity of blossoms as a result of the abnormally large crop produced the previous season. Trees in background are in full bloom. Compare with fig. 10. 228 UNIVERSITY OF CALIFORNIA EXPERIMENT STATION CO rH OS rl w o J I-H 5 w o w Ch Ph o o o P-! Ph w EH «1 OS T— I i-H 02 H 3rH,rH Sh o>" 0, £' b2 "5 CO os CO GO CO O i-H 00 fl a- ; CI O h- rH l^ £5 N CI ° S FH I- rn o cs § (M 00 LO CO LO lO CO tJH LO HH O CO o ^ +^ OJ fl M co a> += . eg CO co£ GO flrrjrH ■H-JS - OS j a> fl rH lo O LO O CXI CO t>^ © o CO b- OS LO CO 2 fi 0) . ^H 00 OS o o o O -~ rH OS LO CO CO to CO M 2^ as o o rH o o o 2 • OS rH i-H oq r-H rH . — O ?h pq « p. -1 CO CD a i fl co CD fl rO HH> .9 © o o CD O • rH rO O ! rH fl CD T \ =H fl CD ^ T3 fH EH rC i O • rH Ph C! i +3 ft CP 02 o3 < Eh f- B rfl ^ O r r c! \ % O rfl CD J cd 4 'S CD • CO I O rC5 r C * H-3 > rH CO o r— i CP O P C CD CO is ■ a ; c ' r2 JO 'y i— 1 OS fl CP fl CD a j T3 T— I ! 13 : a ! r 3 A rH ft fc jd fl OS : fl co : a O CO H H Eh W a r a r a 5 fl H J .. - CP 3 cp i-H fl • rH i — O i ° J © CD i—H Oj i — o ! a 1 K . c 1 r/ 3 ^ fl •r-l CD r- o CD CD ■ rH > CO rH OS M « a a r 1 rf CP i -4J 3 fl CD CD i— * r~ 1 f 2 i * 3 g fl I-H CP 1 — p * f- 1 CD fl 1 fl * rH rH fl •rH r 3 S CP M CP fl 2 fH H-J a 2 P i. ft CD W . ^ rfl r ^ Ph r 1 Ph p — 1 r 1 - H 1 1 P J rH CD cp C ) C3 c 5 3 c3 CD c 1 ^ fl t> fl H ^H 5 pj ^ CD CP 'i 1 rfl 5 fl "- *r 5n ■+■ CD -r- Ph ^ n H -r-l H r-i 3 CD in ft CD 4- CD ' ft r-H CD rH Ph rH OS rH r H rH rH c 1 r - & h Ph Ph 1- H KH M THE COMMON HONEY BEE IN PRUNE POLLINATION 229 . h * • 4 ■ '--Vt > s&$f x ' *£^*#'Sr fl-f ■ ■ ■"■X. / -' -^r •*. ****** Fig. 10. — French prune tree caused to produce a very light crop the previous season by the enforced exclusion of pollen-carrying insects. Note the abundance of bloom. Compare with fig. 9. 230 UNIVERSITY OF CALIFORNIA EXPERIMENT STATION 1915 the French set 5.5 per cent on trees adjoining the Imperials and 2.6 per cent on trees five rows away. In 1916 the results were 3.6 per cent and 3.1 per cent respectively. The decisive difference between 15.1 per cent and 10.8 per cent was obtained in 1917 when counts were made on the same trees as in the previous season. DISCUSSION OF RESULTS The work of the year just passed again shows emphatically the value of bees as pollen-carrying agents. Reports from a number of growers who kept bees in their orchards during the blossoming season, agree with the results obtained experimentally on the Sorosis Ranch. In every case a substantial increase in the size of the crop was re- ported. While no experimental counts were made in these orchards, an inspection of them substantiated the claims put forth by the owners. At the Sorosis Ranch, the increase on the French trees due to the use of bees was remarkable. These insects are at least partly re- sponsible for the increase from 3.6 per cent in 1916, when practically no bees were present, to 13.2 per cent when there were 115 colonies in a 180-acre orchard (Fig. 7). The value of these insects is further shown by the set of 15.5 per cent on French trees adjoining the Imperials and incidentally closer to a number of colonies of bees, as compared with the set of 10.8 per cent five rows farther away from the Imperials and from the bees. Even more striking results were obtained on the French trees which were under the tents. Here the bees and similar insects were entirely excluded or confined, and in the latter case forced to work several days on only the blossoms of the trees under experiment. The trees from which the bees were excluded, though given the same cultural treatment as the check trees, failed to set even a fair crop. The data from both years point to the necessity of some agency for distributing pollen and show the beneficial results from the use of such an agency. The trees themselves produced a vigorous new wood growth (Fig. 12). On the other hand, by confining bees to two trees the percentage of set was greatly increased. Although the marked increase obtained in the experiment could not be obtained in a commercial orchard because of the impossibility of confining bees to a certain section, still a noticeable benefit should result from the keeping of bees in conjunction with the prune orchard. In addition to the data which show that the French prune can be cross-pollinated with the Imperial, the former variety is also appar- ently self -fertile. Under a tent enclosing a single French tree with THE COMMON HONEY BEE IN PRUNE POLLINATION 231 a colony of bees, the percentage of set on the tree was higher than the average for the orchard, and also higher than the set on the tree where cross-pollination with the Imperial was possible. The 19.4 per cent set obtained in this case would clearly seem to indicate that /^r^nc/f Prunes /?/7 Y/////A h.Vtf <>efor?c/er OrcAart/ co/7r/a/ firun^s. ZZ. <7* Average o/-c/iarc/ . W6 ■ )/?# Set cnc/oscc/ *s/M 6./% &** unJer Orc/harJ cono'fr/'ons. Fig. 11. — Production of trees caused to bear heavy or light crops in 1916, com- pared with the production on the same trees in 1917. this variety was amply able to set fruit with its own pollen, providing the pollen was properly distributed. The difference in percentage of set between the French alone and the one enclosed with the Imperial is but 3.9 per cent. It is not improbable that this difference is due to individual tree variation. Neither can these results be construed /9/t BBBH aeaa^^i 232 UNIVERSITY OF CALIFORNIA EXPERIMENT STATION to mean that it is not beneficial to plant Imperials among the French. On the contrary, the results of three years show a higher percentage of set on the French trees close to Imperials than on those farther away. The large yield obtained by using bees on a single French tree, showing a degree of self-fertility, may throw some light on the question why the prune orchards of the state which are largely of this variety have been able to produce fairly satisfactory crops year after year. :"^£ ■"2 f .:"■ '.' V rtm'*taa? Fig. 12. — French prune tree from which all pollen-carrying insects were excluded. It bore practically no fruit but produced a vigorous new wood growth as shown by the shoots near the top of the tree. With the Imperials, the factor of the condition of the trees in all probability had some influence on the percentage of fruit set (Fig. 13). These trees were top-worked on Robe de Sergeant in 1912. In many cases the wounds made in grafting permitted the entrance of heart-rot fungi, with the result that many branches died or broke down each year (Fig. 13). Although apparently vigorous, many of them seemed to be incapable of response to the stimulation of increased pollination. It is noteworthy that the percentage of set was maintained at the high figure of 7.2. THE COMMON HONEY BEE IN PRUNE POLLINATION 233 The Imperial trees kept under tents during the blossoming season also showed decisive results, although somewhat less pronounced than those obtained with the French. With this variety the lack of proper pollen distribution was shown by the lack of fruit in 1916 and the very small (0.43 per cent) set in 1917, when all pollen-carrying agencies were excluded. When enclosed in a tent with a French tree and a colony of bees, the Imperial tree set 7.9 per cent, which was 0.7 per cent above the Fig. 13. — Imperial prune tree broken down under heavy load due to heart-rot fungi which gained entrance through large grafting wounds. average for the orchard. This small increase may have been due to the condition of the tree. While still sound to the casual observer, there was evidence of heart-rot present. Although no individual tree records had been kept during the previous years, the group in which this tree stood had, according to the owner, been producing heavily, and therefore, it might not have been able to respond as fully as other trees in the orchard. Nevertheless the increase, though slight, was consistent with the results obtained with the French variety. It showed, furthermore, that the results obtained in a similar experiment 234 UNIVERSITY OF CALIFORNIA EXPERIMENT STATION in 1916, when the Imperial under conditions identical with those de- scribed, set a much smaller percentage of fruit, were unusual and would not necessarily be a frequent occurrence under commercial orchard conditions. In other words, an Imperial orchard might ordinarily be expected to set a higher percentage of fruit when a large number of bees were present than when they were not. The surprising feature of the work of the 1917 season was the set of 3.02 per cent obtained on an Imperial tree enclosed under a tent alone with a colony of bees. Results obtained up to the present season, in another part of the pollination project carried on by this Division, had indicated that the Imperial was unable to set fruit with its own pollen. In this case, however, after the bees had thoroughly worked on the blossoms for six days, enough blossoms were stimulated into bearing fruit to set a satisfactory commercial crop. It is pos- sible that the Imperial, though usually self-sterile, in occasional years at least is partially self-fertile. It is probable that this self- fertility does not occur very frequently. Similar conditions are found in the case of the apple. 3 The behavior of the trees experimented upon in 1916 presented several interesting features the past season. The French tree which in 1916 was induced to bear an abnormally heavy crop, this year set a light one. Moreover, this difference between 18.05 per cent and 7.4 per cent does not tell the entire story. In 1917 the tree not only had fewer blossoms but actually set a lower percentage of this reduced number of blossoms. Incidentally the blossoms appeared about a week later than the average. They were also fewer in num- ber and apparently weaker both in general appearance and in performance. The entire tree showed the effects of the previous season's overload and made but little vegetative growth. The per- formance of this individual tree may be taken as an example of what may possibly happen in an entire orchard, to a greater or less degree, if the trees are overloaded by suddenly increased pollination. This condition may also last until the trees become adapted to the changed factor of pollen distribution. However, it is not likely that enough bees could be put in the orchard to duplicate the conditions under the tents in the experiments just described. Under the opposite condition where bees were kept from the French tree and a very small crop was set (Fig. 11) in 1916, the yield in 1917 showed a percentage of set of 10.1 per cent. This tree s According to E. J. Kraus of the Oregon Experiment Station, ' ' some varieties of the apple have been found to be self -sterile three years out of five and self- fertile the other two." THE COMMON HONEY BEE IN PRUNE POLLINATION 235 behaved normally in every respect. It produced an abundance of flowers and set an entirely satisfactory crop. Its performance may be likened to what would occur in case a number of colonies of bees were placed in an orchard that without bees had only produced small crops. Both of the Imperial trees used in the experiments in 1916 set but very small crops that year (Fig. 11). When exposed to open orchard conditions in 1917, they both responded by producing crops that were above the average for the variety. The Imperial was thus also shown to be capable of producing larger crops when the number of pollen-carrying insects was increased The trees which bore no crop in 1916 showed a larger wood growth than the average for the variety. SUMMARY The summary of the year 's work is as follows : 1. The average orchard set of French prunes was increased by the use of bees in the orchard. 2. No increase in the average orchard set of the Imperials resulted when bees were used in the orchard. 3. The percentage of set in both French and Imperial prunes was greatly reduced when all pollen-carrying insects were kept away from the trees. 4. The percentage of set on the French tree enclosed in a tent with an Imperial tree and a colony of bees was higher than the orchard average for the variety. 5. The French tree enclosed alone under a tent with bees also set a higher percentage of fruit than the average for the variety. 6. The percentage of set on the Imperial tree enclosed in a tent with a French tree and a colony of bees was slightly higher than the average for the variety. 7. A fair crop of fruit was set on the Imperial tree enclosed alone under a tent with bees. 8. The French tree caused to bear a heavy crop in 1916 by the use of bees set a very small crop in 1917. 9. The French tree caused to bear a light crop in 1916 by the exclusion of bees, set a satisfactory percentage of fruit, although not as high as the orchard average, in 1917. 10. Both Imperial trees which produced light crops in 1916, set heavy crops in 1917 when under open orchard conditions. 11. The percentage of set on the French trees was inversely pro- portional to the distance of the French from the Imperials. 236 UNIVERSITY OF CALIFORNIA EXPERIMENT STATION CONCLUSIONS The results of the two seasons' work seem to warrant the following conclusions : 1. Both the French and Imperial prunes may be aided in setting fruit by the use of bees in the orchard during the blossoming period, provided the trees are in a normal, healthy condition. 2. The absence of bees in the orchard may mean a low percentage of set with both of these varieties. 3. The French prune does not absolutely require interplanting with the Imperial, even though this arrangement may prove beneficial to both varieties.