at” A STUDY iN BLUEBERRY POLLINATION THESIS FOR THE DEGREE OF M. S. T. A. Merrill 1933 t. ‘ WWIWUIUWW WIN! ‘lele ~ ~ _ 31 923 00685 4958 A STUDY IN BLUEBERRY POLLINATION Submi;ted to the Graduate Faculty of Michigan State College of Agriculture and Applied Jcience in Partial Fulfillment of the Requirements for the Degree of Master of Science by T. A. merrill February 27, 1933 :{hwiU I 1. A Study in Blueberry Pollination Introduction The success attending recent propagation studies of the blueberry seems likely to bring this plant into commercial cultivation and thereby to raise other problems, including pollination. Since information available concerning pollination requirements of the varieties which seem most promising for cultivation is very limited, this investigation was planned to secure more information on this point. Review of Literature Concerning Vaccinium blossoms in general Knuth (1) states: "@hey arélfeebly protandrous bee flowers, or flowers with concealed nectar, ****. There are appendages to the anthers, projecting laterally or vertically, and when insect visitors strike against these the pollen is shaken out. *** "Kerner says that in the species of Vaccinium automatic self-pollination is ultimately possible, for the flowers, though erect at first, become pendulous, so that pollen can fall on the stigma." After this general statement concerning Vaccinium, Knuth summarizes his own observations and those of others for four species: V. myrtillus, V. uliginosum, V. Vitis- Idaea and V. oxycoccos. is a whole, these indicate (1) 102685 2. prevalence of homogamy, with slight tendencies toward protandry; (2) adaptation to insect visits, which in some cases insure crossing; (3) automatic self-polli- nation in some cases and lack of information in others; (4) cleistogamy or pseudocleistogamy in two cases; (5) in most cases a fair number of insect visitors, among which the bumble bee is prominent. No observations on V. corymbosum are offered. Concerning this species Coville (2) states: "When blueberry flowers are pollinated with pollen from their own bush the berries are fewer, smaller, and later in maturing than when the pollen comes from another bush. Some bushes are almost completely sterile to their own pollen. *** It is important, therefore, that a plantation should not be made up wholly from cuttings from one bush." Beckwith (3) states: "Close study has shown ** that self-pollinated blueberries are small and late in matur- ing even when the pollen was carried promptly from flower to flower by bumble bees. Berries of commercial size and quantity could not be produced in this way. Cross polli- nation is accomplished by means of wild bees in the usual small patch but in the larger plantations, colonies of honey bees are introduced and are very active during the main blooming period. It has been noted that ** during the main and late blooming periods, they (honey bees) are the chief insects in the fields. Bumble bees and carpenter bees as well as other wild bees are common throughout the blooming period. *** Two insect proof 3. cages were built, each covered two bushes of different varieties and in one a small colony of honey bees were introduced. There was a good set of berries in the cage with the honey bees and no set in the insect-free cage, indicating that honey bees unassisted are able to pollinate blueberries satisfactorily." Coville and Beckwith are quoted at some length here because the findings of the study now reported differ considerably from their views. Both of these writers are clearly of the opinion that blueberries are self sterile. Beckwith's eXperiment, it will be noted, shows, as reported, merely that insect agency is necessary to pollination. Coville (4) reports some experimental work tending to indicate self-sterility, but his report is so meagre in details of technique employed as to leave the conclusions Open to some doubt. methods The field investigations here reported were carried on at the South Haven Branch Experiment Station in 1952. Previous pollination study on blueberries was almost wholly lacking and methods previously outlined were necessarily modified as the work progressed. Bagging: Small branches containing from 100 to 200 blossoms were covered with 5 lb. paper sacks. A few muslin bags were used as checks against the paper bags. Five cultivated varieties and the wild highbush and wild lowbush were thus treated. At least 24 bags were placed 4. on each variety used and in the wild highbush 36 bags were used, 27 paper and nine muslin. All open flowers on the branches used were picked before the sacks were applied. The bags were left over the branches throughout the blooming season (two weeks). The purpose of the bags was two fold: first, to keep all insects from the blossoms, and second, to determine whether those blossoms would be pollinated by their own pollen and set fruit. (Fig. I). Caging: Eighteen cages of l4—mesh screen and of size ample enough to cover the bushes without crowding were set securely over the plants selected for this study. The screen extended several inches below the soil surface and the cages were tight against insects larger than thrips. As a check against these, a single 3' x 3' x 4' nainsook cage was used. To guard against thrips and other accidental small insect visitors, this cage was fumigated with carbon bisulphide for three Consecutive days and after a one-day interval for two days more, thus covering practically the whole blossoming season for the enclosed bush. The screen cages were set in place before the blossoms opened and those blossoms that were open when the nainsook cage was set in place were removed. (Figs. II, IIa, III and IIIa). Emasculation: The blossoms should be emasculated just before they open., Inasmuch as there is much variation in degree of blossom development of blossoms in any cluster all open blossoms and all that were too small were removed. It was impossible to employ the methods of emasculation 5. that are commonly used for the orchard fruits. This is due to the difference in the structure of the flower. The tissue of the blueberry blossom is much more tender than that of the tree fruits, and the flower separates 'from the stem very easily. This tenderness necessitates great care in manipulation. The flowers were emasculated with greatest speed and accuracy by means of small eyebrow tweezers filed down to rather sharp points. The stem of the flower was held with one hand while the other hand brought the point of the tweezers to the base of the blossom and exerted just enough pressure to make the point of the tweezers barely pierce the petal. Then the points are pressed together and an upward movement removes a small portion of the united petals. Two to four trials are necessary to remove all the petals and stamens. The pistil is by far the most tender part of the flower and a very slight touch breaks it off at the base. The slightest pinch or bruise from the tweezers causes the upper part of the pistil to wither and renderes it useless for further work. In all cases pollen was applied at the time of emasculation; thus bagging of the emasculated blossoms was not necessary. This, however, was checked by emas- culating several clusters of blossoms and leaving them without pollination. (Fig. IV). Collecting_the Pollen: After it was found impossible to force the blossoms and collect the pollen in advance, as is usual in studies on apples, etc., it was discovered that by picking the opening blossoms and holding a blossom 6. in an inverted position between the thumb and first two fingers and rolling lightly, the anthers were broken and the pollen forced out. This method was used throughout the period of pollination. The pollen, as it was forced from the anthers, was caught in a petri dish fitted with a cover. Since a very few minutes sufficed to collect enough pollen for a cross, collections were not made until time for its use arrived. Thus it was certain that only fresh pollen was being used. Some question may be raised as to the possibility of contamination, since partly cpen blossoms were used. Observation of Opening blossoms and bee preferences indi- cated that this danger was very slight. Certainly no contamination wflhhithe bounds of reasonable assumption could account for the results to be reported. gpplication of Pollen: In all cases the pollen was applied immediately after emasculation. Application to the stigmas was made with a finger. This method was more rapid and gave a more complete coverage of the stigmas than was secured with a camel's hair brush. Frequent sterilization of the finger in absolute alcohol obviated possibility of mixing of pollen. The pollinated pistils were not covered in any case, but as a check onthe possi- bility of natural pollination after the artificial polli- nation 146 flowers were emasculated on one variety and left unpollinated and exposed. (Fig. IV). Three weeks later there was no indication whatever of a single berry developing. 7. Pistils that were damaged, even very slightly, wilted and turned brown almost immediately. Inasmuch as from an hour and a half to two hours were needed for emasculating enough blossoms for a single cross, (from 125 to 200), it was very easy to detect and eliminate any damaged pistils. These rarely exceeded Sflfw’”fl Three or four weeks after pollination a count was made of the apparently developing berries. In some cases it was impossible to draw any sharp line between the berries that would drop and those which would continue development. These counts, however, were recorded. Germination: In this test several germination media were used which will be discussed later in this report. Petri dishes, containing 10 c.c. of each medium, were used as germinating chambers. After the germinating medium was placed in the petri dish a few pollen grains were added and the dish was then fitted with a cover and allowed to stand for 24 hours. In making the counts only those grains were counted as germinated which had sent out a tube. 8. Tab 16 I Showing weather conditions — May 12 to 27 Temperature ‘ Character Date MaXimum Minimum, Precipitation of Dgy 12 56 48 Partly cloudy 13 76 34 " " 14 85 48 Clear 15 87 50 " 16 80 48 .05 Partly cloudy 17 64 30 Clear 18 68 28 Partly cloudy 19 74 38 " " 20 8O 38 Clear 21 83 44 Partly cloudy 22 75 37 ” " 23 73 28 Clear 24 87 37 Clear 25. 86 54 .10 Partly cloudy 26 '75 57 1.65 Cloudy 27 63 56 Partly cloudy PRESENTATION OF RESULTS Germinability of Pollen Germination tests were made about the middle of the pollinating season. The percentages, as appear in Table II, were so high that there is reason to believe that the pollen was in good condition throughout the season. Since, as has been explained, the blossoms could not be forced to maturity by artificial means, all the pollen used came 9. directly from blossoms that matured on the bush. Table II. Germinability, in cane sugar solution, of pollen used Commnhnttn cfcnmaswmr 3% 6% 9% 12% sdhmhxi Gnmms 3% (Reins %§ Gunns 9% (henna % Variety counted m counted germ. counted gem. counted germ. Rubel 19 26 28 42 77 83 86 11 Cabot 36 19 39 12 31 48 21 90 Adams 13 38 28 80 41 92 53 100 Pioneer 44 52 46 65 84 84 60 80 Harding 37 13 32 12 55 50 31 50 The germination tests show rather consistently better results with the higher concentrations of the medium. Among the varieties, Cabot and Harding showed lowest germination. Since only one test Was made, however, no great significance is attached to these differences, particularly in view of the satisfactory sets obtained when pollen of these varieties was applied to developing pistils. Determination of Normal Set of Fruit for Blueberries To determine the normal set of fruit among the varieties used, several branches of each variety were selected at random and their blossoms counted and tagged. This first count was made May 16. When the blooming season is over, practically every flower is succeeded by a miniature blueberry. Unlike the tree fruits, the blueberry drop is very light, providing the season is at all normal. 10. dven though part of these miniature berries do not continue growth they remain green and persist on the bushes for three or four weeks; at this time they dry and fall, though a few may persist to the time of picking. It is necessary that more of these small fruits reach maturity in order to give a full crop than is the case with tree fruits. Just what really determines which of these very small fruits shall drop is uncertain. It is prObable that for the greater part it is due to the lack of fertilization of the ovules. After the first drop by far the greater part of the remaining berries may be expected to reach maturity, pro- vided weather conditions are favorable. he number of flowers which develOp into mature fruit depends, theoretically at least, on a number of factors: (1) fertilization of the ovules, (2) the age, vigor and general conditions of the bush, (3) the available nutrient and moisture supply in the soil, (4) the relative heaviness or lightness of the bloom. Mr. Stanley Johnston's observations based on several years eXperience with these varieties indicate that a set of from 85% to 90% or eight or nine fruits from every ten blossoms is necessary for a good crop. The estimated percentages of full crops borne on all varieties are the averages of estimates made by Mr. Stanley Johnston and by the writer, except for Harding which was estimated only by Mr. Jehnston. These estimates were made at and during the time of harvest and were to some extent based on comparison with previous seasons' crops. 11. The oldest plants, which were planted in 1927, were the only bushes considered when these estimates were made. As previously stated, the branches chosen from which the normal set was computed were selected at random but were average branches and open to insect pollination. In a1 cases several bushes were used. 1 In the wild highbush and wild lowbush no estimate of the per cent of full crop borne was made. Table III. Normal set of fruit among blueberry varieties open pollinated. Flowers Apparent se’tjar Harvest . Variety counted Date Numb er % berries May 16 June 15 (July) of berries matured Rubel 991 89.7 15 847 85.4 Cabot 931 58.1 12 540 58.0 Adams 922 89.3 13 820 88.9 Pioneer 624 91.3 13 570 91.3 Harding 610 60.3 13 364 59.6 Wild lowbush 676 -- 13 92 13.9 Wild highbush* 1850 -- 12 849 45.6 * Observations made at Bangor, Michigan. Perhaps the most remarkable feature of these data is the slight amount of drop in every case. percentage of drop rather accurate early crop estimates could be made after the first drop, providing weather conditions are favorable. In hand-pollinated blossoms (cf. Table VII) the late drop is considerably greater. Due to this low Self-Fruitfulness The methods followed in self-pollination were essentially the same as those previously described, except that three distinct methods were used; two of 12. these yielded results shown in Tables IV and V and the third is illustrated by Fig. 5. In one case the flowers were bagged with five pound paper bags which were placed over the blossoms prior to opening and left until the blooming season was over (Fig. I). In the second case the blossoms were emasculatéd and pollinated by hand, as previously described. In the third case the plants of one variety, Rubel, were secluded from all other varieties and left for insect pollination. sented in Tables IV, V, and Fig. 5. Results of self-pollination. Table IV. (No pollen applied) Flowers bagged. The results are pre- Bagggl Harvest Variety Date No. No. wars Dfie No. (May) sacks enclosed only) berries natured Rubel 19 24 3028 15 1365 45.1 Cabot 13 20 4349 12 375 8.6 Adams 16 22 2920 13 930 31.8 Pioneer 16 20 3761 13 185 4.9 Harding 16 24 3782 13 70 1.9 WildIbIbmh 14 20 2720 14 0 0 was 1113mm: * 12 27 4439 12 o o * Test made at Bangor, Michigan. 15. Table V. Results of self pollination (Hand pollination, May 20) 11 gm arantset Harvest ( Variety No. Jume 18 Date No. % fruits pollinated No. 3' (July) picked matured Rubel 123 115 93.5 15 104 84.5 Cabot 99 90 90.9 12 80 80.8 Adams 131 115 87.7 13 107 81.6 Pioneer 118 108 91.5 13 108 91.5 There is a manifest difference in the set secured under sacks and that outside them. The fact that these blossoms were enclosed in sacks, while the hand pollinated blossoms were not, suggest that heat, humidity or light conditions under the sacks were such as to cause injury to blossoms, to prevent proper dehiscence of the anthers or to prevent fertilization or to cause unnatural stimulus to the formation of the absciéion layer. It is possible also that the sacks reduced atmospheric and mechanical movement to such an extent that pollen did not readily reach the stigmas, which protrude considerably beyond the anthers. A trial already reported (page 6) shows that omission of the sacks did not vitiate the reliability of the data recorded in Table V. The mechanical aid as discussed above must therefore be left to insects, principally bumble bees and honey bees and it seems probable that they play an important role even in self pollination. Figure 5 illustrates the role that bumble bees and honey bees play in the pollination of blueberries. It shows 14. 48 Rubel plants in the third season of growth in the field. These plants are at the EXperiment Station Grounds at South Haven and are three miles from any other blueberry plants. Practically~speaking it was an impossibility for pollen to be carried three miles by bumble bees and honey bees and no honey bees were placed in the plantation three miles south during the blooming season. These plants set a very satisfactory crop for the third season - 25.77 pounds of berries were produced. The writer was particularly careful to observe the bumble bees and the honey bees working on these plants during the blooming period. To compare the per cent of full crop given by self- pollination in relation to the normal set for the season, Table VI has been compiled from Tables III and V. Table VI. Showing per cent of a full crop given by self-pollination. Normal set (73 full crop 95 from self- % estimat Variety for season borne pollination mchrop by saflhpiui- nation Rubel ' 85.4 100 84.5 98.9 Cabot 58.0 66 80.8 91.9 Adams 88.9 100 81.6 92.9 Pioneer 91.3 95 91.5 95.2 These data show that Cabot and Pioneer both gave a higher percentage of set when self-pollinated than from open insect pollination, the difference being considerable in Cabot. Adams would undoubtedly have shown as high figures as the open pollinated had not the plant on which 15. the selfing was done been affected by the mummy disease which caused many of the berries to fall before they were first counted. In the case of Rubel the difference is very slight and a good crOp could be assured from a set as good as is indicated above. Inter-Fruitfulness The pollination work carried out by Coville with blue- berries indicated that the blueberries with which he worked were nelf-unfruitful but were highly inter-fruitful. Because of the commercial importance of Rubel the greatest number of crosses were made on this variety. (Table VII). The data show that the percentages of set, as a rule, ran rather low and very low in some cases. With one exception, Harding pollen produced the highest percentage of set. In the majority of cases the percentage of set is too low to give a good commercial crop. The differences between the percentages for June 18 and those for July 12 - 15 represent the percentage of flowers pollinated which dropped after June 18. The drop in most cases is great enough to be of some impor- tance, and of considerable importance in a few cases, in affecting the size of the crop borne. Table VII. Results of cross-pollination. | Pollination gamut set Harvest Cross D913 No. uhe 18 Date 0 dew) blnsoms 1N0. ‘% (Jun) No. :mahmed hube1 x Cabot 19 216 193 94.7 15 155 71.6 7 1 ideas 20 227 219 96.4 15 190 93.7 " 1 Pioneer 19 208 181 87.0 15 169 81.2 n x Harding 20 212 212 100.0 15 192 90.5 Cabot x Rubel 17 127 95 74.8 12, ' 81 63.7 " x idams 17 121 99 91.0 12 74 61.1 n x Pioneer 17 112 92 73.2 12 i 72 64-0 7 x I-Brding 17 105 95 91.0 12 69 95-7 (Adams 1: Rubel 18 170 157 92.3 13 126 74.1 v x Cabot 19 157 144 91.7 13 129 80.2 " x Pioneer 18 125 117 93.6 13 98 78.4 n 3: Holding 19 126 112 99.9 13 '78 61-9 Pioneer x Rubel 17 130 121 93.0 13 115 88-4 n x Cabot 17 111 93 93.7 , 13 ' 93 93.7' " x Adams 17 114 59 51.7 13 58 50.8 7 x Harding 17 127 119 93.4 13 109 95.0 Harding 2: Rubel 19 122 114 93.4 13 91 74-5 a x Cabot 19 90. 64 71.0 13 48 53-5 n 1: Adams 19 109 100 91.6 13 '78 72.2 " 1 Pioneer 19 85 85 100.0 13 66 77.6 Table VIII. 17. Comparison of set with self-pollination and with cross pollination. Self- Variety pollination Cross pollination LTableVIfl (Tgble V) Cross Per cent Rubel 84.5 Cabot 71.6 Adams 83.7 Pioneer 81.2 Harding 90.5 Cabot 80.8 Rubel 63.7 Adams 61.1 Pioneer 64.0 Harding 65.7 Adams 81.6 Rubel 74.1 Cabot 80.2 Pioneer 78.4 Harding 61.9 Pioneer 91.5 Rubel 88.4 Cabot 83.7 Adams 50.8 Harding 85.0 Harding -- Rubel 74.5 Cabot 53.3 Adams 72.2 Pioneer 77.6 18. Self-pollination with one exception gave higher set than cross pollination. (Table VIII). In most cases the individual differences are not great enough to establish significance, but the consistency of the direction of the differences is rather re- markable. Since crossing was done on May 17, 18, 19 and selfing was done on hay 20, there is a possibility of selfing having been done when conditions, internal or external, were most favorable. Weather records, (Table I) however, do not show one time to have been more favorable than another. Technique was uniform throughout. Crosses were not made until the stage of development seemed to require that the work be done. The differences in results secured may have been due to a mistaken notion of the proper stage, so that the cross pollination was somewhat premature. The data reported in Table XI tend to support this explanation. Other than this no suggestion can be advanced now. Certainly it is not Claimed that self-pollination is uniformly superior to cross-pollination. The practical con- sideration, that self-pollination is adequate, seems, however, well established. Selfed berries were as large in every case as the crossed berries. There is no reason whatever to believe that self-pollinated blueberries did not develop to full normal size. The mean diameter (in cm.) of 86 Rubel berries from self-pollination (by hand) was 1.517ltTO.2147 and of 100 Rubel from open pollination, in the same picking, 19. 1.375IIO.2023. NO relationship appears between size of berry and number of seeds, the correlation found on 400 Open-pollinated Rubel berries being 0.0332 t0.0336. The results from self-pollination here recorded differ so from previous reports that they have been re- examined with great care and any conceivable eXplanation sought. NO possibility Of error great enough to affect the results materially has been found. The relative value of the various cross-pollinations, as compared with the crop borne on the remainder of the bush and Open to insect pollination is more clearly shown in Table IX. Table IX. Comparison Of percentages Of full crop produced by pollen of various kinds with those pro- duced with Open pollination. Open pollination Hand pollination Female Male Normal set Estimated % 73 a parent parent for season full crOp fruits estimated % bomna mahmng :fldltrop Rubel -- 85.4 100.0 7 Cabot ’ 71.6 93.9 " Adams ' 93.7 99.0 " Pioneer 81.2 95.0 " Harding 90.5 105.9 Cabot -- 58.0 66.0 Rubel 63.7 72.5 Table IX (Cont'd) Open pollination Hand pollination Female Male Norma]: set Estimated % 7,; K parent parent fir ocean full crOp frui estimaied % borne naturhg full cmp‘ Cabot Adams 61.1 69.5 " Pioneer 64.0 72.8 " Harding 65.7 74.7 Adams -- 88.9 100 " Rubel 74.1 82.2 " ’ Cabot 80.2 88.9 " Pioneer 78.4 87.0 " Harding 61.9 69.6 Pioneer -- 91.5 95.0 ” Rubel 88.4 91.9 " Cabot 83.7 87.1 " Adams 50.8 52.8 " Harding 85.0 88.4 Harding —- 59.6 75.0 I " Rubel 74.5 93.8 " Cabot 53.3 67.1 " Adams 72.2 90.9 " Pioneer 77.6 97.7 4 Column two Of Table IX gives the estimated per cent of a full crop borne by the bush under ordinary conditions and the last column gives the per cent of a full crop pro- duced by the cross. In nine out Of 20 crosses the per- centage of set was higher than the normal set for the season. This condition is believed to be due to the lack 21. of visits by insects to the "Open pollinated" blossoms that were included in the counts for the normal set for the season. No one variety of pollen gave the highest percentage of set on all varieties. Fruitfulness in Reciprocal Crosses As a matter of convenience and interest, crosses already reported (Table VII), are grouped as re— ciprocals in Table X. Here the consistently poor record of Cabot as a female parent, evident in Open pollination and in controlled crosses, stands in con- trast to its rather high performance as a male parent. This, in conjunction with its satisfactory performance when artificially self-pollinated, suggests structural peculiarity or partial dichogamy as explanation Of the rather poor results with Open pollination. There is, however, no reason to expect better results from inter- planting for cross pollination Of this variety. Table X. Results from reciprocal crosses. I Per cent ##Per centIset of (Cross set reciprocal cross Rubel x Cabot 71.6 63.7 Rubel x xdams 83.7 74.1 Rubel x Pioneer 81.2 88.4 Rubel x Harding 90.5 74.5 Cabot x Adams 61.1 80.2 Cabot x Pioneer 64.0 83.7 Cabot x Harding 65.7 53.3 Adams x Pioneer 78.4 50.8 Adams x Harding 61.9 72.2 Pioneer x Harding 85.0 77.6 Duration of Pistil Receptivity 22. To determine the length of the period of pistil receptivity several hundred Rubel flowers were emasculated on May 21 and pollinated at intervals, ranging from immediate application tO 96 hours. results of this test are presented in Table XI. The Table Receptivity Of Rubel pistils to mixed pollen. Set Set Hours to NO. blossoms June 18 July15 Harvest July 25 pollination pollinated % % No. a 0 116 81.0 71.5 83 71.5 10 141 92.9 83.6 118 83.6 24 97 89.6 88.6_ 86 88.6 49 159 75.4 66.0 105 66.0 72 63 100.0 95.2 60 95.2 96 66 96.9 95.4 63 95.4 The striking feature of the data in Table XI is that blueberry pistils are more receptive after 96 hours than they are immediately or even ten to 24 hours after emas- culation. As already mentioned, this may have some bear- ing on the superior set obtained with self-pollination. This behavior was so unexpected that no provision had been made for continuance of the test beyond 96 hours; consequently the optimum and the maximum periods cannot be stated. Since these blossoms were not sacked, it was foreseen that lapse of time would increase chance of insect visit 23. with resultant pollination. To check this possibility 146 blossoms on a neighboring bush were emasculated and left for chance pollination. Of these, six set and none matured (Fig. 4). Consequently, the conclusion as to duration of receptivity seems entirely warranted. Notes on Pistil Structure Inasmuch as no discussion of the anatomy Of the blueberry blossom has been found in the literature consulted, photomicrographs showing its general structure are presented (Figs. IX, X, XI, X11 and XIII). Samples of blossoms collected at intervals up to 72 hours after pollination showed, when sectioned, no tube growth beyond the stage shown in Fig. XIII, which was found only in the 72-hour material, though pollen grains were abundant in other material. Accordingly, no attempt is made here to discuss the structural adaptations Of the blossom to fertilization. The material shows clearly that uninterrupted open channels lcad directly from the stigma to the ovarian cavity, but it has not been demonstrated that the pollen tubes follow these channels. Summary 1. The blueberry was found to be self-fruitful. 2. NO better results were obtained with cross pollination than with self pollination. 3. The pollen germination tests showed rather consistently better results with the higher 24. concentrations of the germinating medium. 4. Blueberry pistils are more receptive after 96 hours than they are immediately or even ten to 24 hours after emasculation. 5. Bumble bees and honey bees are the principal pollinating insects. acknowledgments The writer wishes to express his appreciation of the services rendered him in this study. Mr. Stanley Johnston Offered some assistance in the pollination work and gave many helpful suggestions; he also did the field photographic work. Professor F. C. Bradford has also given advice and criticism. Thanks are due also to W. C. Dutton for his laboratory photographic work and to Director V. R. Gardner, Dr. R. E. Marshall and Dr. J. W. Grist for the many helpful suggestions and advice offered. Literature Cited (1) Knuth, Paul Handbook Of Flower Pollination III: 27. Oxford, 1909. (2) Coville, F. V. Directions for Blueberry Culture, 1921, U. S. D. a. Bul. 974. 1921. (3) Beckwith, C. S. Rept. Dept. Ent. N. J. State Agr. EXp. Sta. 1930. p.174. (4) Coville, F. V. . EXperiments in Blueberry Culture. U. S. D. A. Bur. Pl. Ind. Bul. 193, 1910. Branches with paper bags I. during blooming season. Fig. . Fig. II. Showing galvanized screen cage over plant at blooming time. 11L, .Fig. II A. This plant was under galvanized screen cage during blooming season. ' Fig. III. Nainsook cage. Fig. VIII- shows some berries taken from this bush. .Fig. III A. Harvest time condition of plant covered with nainsook cage during blooming. -Fig. IV. Branch in which blossoms were emasculated and left for open pollination. Six berries set and none matured. . 9? an; (_ '. -_ .9339 d .Fig. V. Shows 48 Rubel plants in the third season. These plants produced 25.77 pounds of berries. They were pollinated, by insects, with their own pollen. 'Fig. VI. Showing branch in which blossoms were emasculated and pollinated (by hand) with own pollen. \ 1 Tormal set on Rubel from open ‘ Fig. VII. pollination. Fig. VIII. Berries picked from bush covered with nainsook cage during blooming period. at left Center: At right: fully normal berries picked Aug. 13. same number Of berries picked same day but green; these berries developed parthenocarpically. ripened parthenocarpic berries, picked Sept. 9. Fig. IX. Longitudinal section of emas- culated blueberry blossom. anther ‘attachment shown at A; portion Of petal at B. Lumina in pistil not shown as continuous because Of bending. Rig. X. Transverse section of upper portion Of blueberry style. Sub- stantially the same condition obtains throughout the style. locules and vascular connection Transverse section of basal portion XI. of pistil showing to the ovules. Fig. ! Showing path Of stylar canal' XII. into the locules. 'Fig 0 Fig. XIII. Germinating pollen grains on pistil, 72 hours after pollinatiOn. 1 I‘ll ‘I. 111129 4 7 W 19 MM; ROOM USE 01111 ICHIGRN STATE UNI I.V I III IIIII IIIIIIIII IIII IIII II IIIII III IIII III IIIII IIIIIII