Cluster Analysis of Characteristics of Different Sweet Cherry Varieties

Xiaohua YANG,Rong YIN,Chuanming PANG,Guowei NIE,Qianru ZHANG,Kai LI,Huiling RU,Yiwen JI,Zemin DUAN

Pomology Institute,Shanxi Academy of Agricultural Sciences,Taigu 030815,China

Sweet cherry is known as"the first branch of spring fruit"[1],and is favored by consumers due to its abundant nutrition,good flavor and taste and various natural components.Currently,the cultivation of sweet cherry in China is mainly concentrated in the BohaiBayarea.Shanxi Province isa new district growing sweet cherry,and after years of exploration and accumulation,sweet cherry industry in Shanxi Province has made considerable progress.The sweet cherry varieties planted in China are mainly European cherry varieties and Chinese cherry varieties.Between them,European cherry varieties have advantages of early maturity,large fruit size,bright color,abundant nutrition,good taste ad high economic efficiency,which are contained by all fresh cherry varieties with commercial value[2-4].As sweet cherry has been cultivated massively,its botanical characteristics and horticultural cultivation characteristics are not only the basis for variety identification[5],but also the basis for pruning,water and fertilizer management and diseases and pests control.At present,the cultivation patterns of sweet cherry include traditional field cultivation and facility cultivation.The common tree types of sweet cherry include heart shape with three scaffold limbs open[6],small canopy sparse layer shape[7],slender spindle shape[8]and freed spindle shape[9].Wei et al.[10]studied the growth status,fruiting branch type,total bud number and fruit economic traits of sweet cherry,and they found that the fruits of Hongdeng and the other five sweet cherry varieties mainly fruited on the bouquet branches.Yang et al.[11]considered that the selection of sweet cherry variety with high fruit setting rate,as well as adequate short fruit branches and bouquet branches,was a prerequisite for high yield of cherry,and in the highyielding sweet cherry plants,the long fruit branch:medium fruit branch:short fruit branch:bouquet branch ratio was about 3 ∶1 ∶8 ∶40.It was obviously in planting practice that most pruning schemes are formulated based on the characteristics of a few varieties,instead of botanical characteristics of different varieties,thus the existing pruning schemes are similar.To deal with this issue,cluster analysis of natural fruit seedling rate,branching rate and fruiting branch composition was conducted among the 14 sweet cherry varieties, including Hongmanao,Hongdeng and Napoleon.Based on the study results,the pruning schemes for different variety groups were adjusted,thereby providing technical basis for pruning and integrated management of sweet cherry.

Materials and Methods

Field survey

The field survey was carried out in the sweet cherry garden of the Pomology Institute,Shanxi Academy of Agricultural Sciences in 2011-2012.The surveyed indicators included fruit setting rate,branching rate,flower bud to leaf bud ratio,long fruit branch proportion,medium fruit branch proportion,short fruit branch proportion,bouquet branch proportion and yield per plant.The surveyed materials included Hongmanao, Rainier, Longguan,Longxu,Hongdeng,8-2,Napoleon,Burlat,Sato Nishiki,1-5,Bing,Black Tartarian,Wannika and 6-19,total 14 sweetcherryvarieties.The sweet cherry plants were all 16 years old.The plant and row spacing was 3 m×4 m.The trees were all improved trunk-shaped,and they were all robust.The routine management was as usual.

Fruit branch typeThe branch with length higher than 15 cm was defined as long fruit branch;the branch with length ranging from 5 to 15 cm was defined as medium fruit branch;the branch with length ranging from 2 to 5 cm (including 5 cm)was defined as short fruit branch;the branch with length less than 2 cm was defined as bouquet branch.

Fruit setting rateThere was only one sweet cherry tree in each plot.The sweet cherry trees were selected randomly.There were 10 repetitions for each indicator.At the flowering stage,the flower numbers on four medium fruit branches in four different orientations of each plant were investigated.The fruitlet numbers were investigated 3-4 weeks after the full bloom stage.And then,the percentage between fruitlet number and flower number on each branch was calculated.The final data were expressed as the means.

Branching rateThere was only one sweet cherry tree in each plot.The sweet cherry trees were selected randomly.There were 10 repetitions for each indicator.At the defoliation period,total 52-year-old branches were sampled randomly outside the canopy of each sweet cherry tree,and on each branch,the numbers of medium and long fruit branches were investigated.And then,the proportions of medium and long fruit branches in total germinated sprouts were calculated.The final data were expressed as the means.

Flower bud to leaf bud ratioThere was only one sweet cherry tree in each plot.The sweet cherry trees were selected randomly.There were 10 repetitions for each indicator.At the germination stage, total 5 1-year-old branches were sampled randomly outside the canopy of each sweet cherry tree,and on each branch,the numbers of flower buds and leaf buds were investigated.And then,the ratios between flower buds and leaf buds were calculated.The final data were expressed as the means.

Proportion of each type of fruit branchesThere was only one sweet cherry tree in each plot.The sweet cherry trees were selected randomly.There were 10 repetitions for each indicator.At the dormancy period,total 3 big branches were sampled randomly outside the canopy of each sweet cherry tree,and on each branch,the numbers of fruit-bearing branches and fruit branches at each type were investigated.And then,the proportion of fruit braches at each type in the total fruit-bearing branches was calculated.The final data were expressed as the means.

Data analysis

Using the DPS 14 software,the means,standard deviations and variation coefficients were calculated,and the cluster analysis and T-test were conducted.Most of the tables and figures were prepared by the DPS 14 software,and the others were prepared by Excel.

Results and Analysis

Statistics and analysis of field survey data

As shown in Table 1,the yield per plant of the 14 sweet cherry varieties ranged from 3.0 to 18.5 kg.Among the 14 sweet cherry varieties,Hongmanao showed the highest yield,while 8-2 showed the lowest yield.The fruit setting rates of the 14 test varieties ranged from 4.0% to 41.2%.Hongmanao also showed the highest fruit setting rate,while Longxu showed the lowest fruit setting rate.The proportions of bouquet branches in the 14 sweet cherry varieties ranged from 20.0% to 84.0%.Hongmanao had the highest proportion of bouquet branches in the total branches,while Burlat had the lowest proportion of bouquet branches.

Cluster analysis of fruit branch type,fruit setting rate and yield among different varieties of sweet cherry

Fig.1 showed that the 14 sweet cherry varieties were divided into two large groups.The first large group included Hongmanao, Longguan,Longxu,Hongdeng,Black Tartarian,Wannika,Burlat and 6-19,while the second large group included Rainier,8-2,Napoleon,Sato Nishiki,Bing and 1-5.

The first large group was further divided into five small groups.Longguan and Longxu were clustered into one group;Hongdeng and 6-19 were clustered into one group;Black Tartarian and Wannika were clustered into one group;Hongmanao and Burlat were clustered into one group.

The second large group was further divided into two small groups.Bing,Sato Nishiki and 8-2 were clustered into one group,while Rainier,Napoleon and 1-5 were clustered into the other group.

Statistics and analysis offruit branch proportions,fruitsetting rates and yields of different varieties of sweet cherry

The fruit setting rate,branching rate,flower bud to leaf but ratio and the other four indicators of the two types of sweet cherry varieties were analyzed.As shown in Fig.2,there were significant differences in all the indicators,except the proportion of bouquet branches,between the two types of sweet cherry varieties.The results of T-tests showed that the differences in yield per plant,branching rate and proportion ofshortfruit branches between the two types all reached an extremely significant level,while in fruiting setting rate and proportion ofmedium fruitbranches reached a significant level(Table 2).Thus,in order to improve the yield of the two types of sweet cherry varieties,the pruning of tree type should focus on improving fruit setting rate and proportion of short fruit branches,and controlling properly the proportions of medium and long fruit branches.

Conclusions

The cluster analysis divided the 14 sweet cherry varieties,including Hongmanao, Rainer, Longguan,Longxu,Hongdeng,8-2,Napoleon,Sato Nishiki,Bing,Black Tartarian,Wannika,Burlat,1-5 and 6-19,into two types.The first type included Hongmanao, Longguan, Longxu,Hongdeng,Black Tartarian,Wannika,Burlat and 6-19,while the second type included Rainier,8-2,Napoleon,Sato Nishiki,Bing and 1-5.

The yield of first-type sweet cherry varieties,represented by Hongmanao,was higher than that of second-type sweet cherry varieties,represented by Rainier.There was significant difference in yield per plant between the two types of sweet cherry varieties(P＜0.01).

The differences in yield per plant,branching rate and proportion of short fruit branches between the two types of sweet cherry varieties reached an extremely significant level,while in fruit setting rate and proportion of medium fruit branches reached a significant level.To improve the yield of secondtype sweet cherry varieties,their fruit setting rates can be improved by increasing the proportion of pollinator tree and other kind of pollinators in production,and their pruning should focus on increasing proportions of short and bouquet fruit branches but controlling properly proportions of medium and long fruit branches.

Discussion

In this study,the proportion of bouquet branches in first-type sweet cherry varieties,represented by Hongmanao,is almost equivalent to that in second-type sweet cherry varieties,represented by Rainier,but the yield of first-type varieties is much higher than that of second-type varieties.It suggests that the key points of improving yield of second-type sweet cherry varieties are to promote the transformation from vegetative growth to reproductive growth and to improve fruit setting rate.In terms of branch type composition,the second-type sweet cherry varieties have stronger branching capability,and the proportion of short fruit branches in second-type varieties is significantly lower than that in firsttype varieties.Therefore,the adoption ofappropriate pruning techniques,such as topping,twisting tip and carving bud,will contribute to the transformation from development branches and medium and long fruit branches to short fruit branches,improving yield of sweet cherry.

In cultivation practice,Hongmanao,Rainier and Bing can pollinate each other,but their pruning targets are significantly different.Hongmanao and Longguan belong to the same sub group.However,the proportion of long fruit branches in Longguan is much higher than that in Hongmanao,but the proportion of bouquet branches is much lower than that in Hongmanao.Therefore,the pruning the Longguan should focus on controlling the proportion of long fruit branches,but increasing the proportion of bouquet branches.In addition,appropriate technical measures should be adopted to improve the fruit setting rate.Longxu,along with Longguan,belongs to the same sub group.The main pruning objectives of Longxu are similar to those of Longguan.However,the fruit setting rate of Longxu is much lower than that of Longguan.Therefore,to improve the yield of Longxu,the tree growth should be controlled,the pro-portion of long fruit branches should be decreased,and the fruit setting rate should be increased.Bing and Hongmanao belong to different sub groups.The flower bud to leaf but ratio of Bing is much higher than those of Longguan and Longxu,and the proportion of bouquet branches in Bing is much lower than that in Hongmanao.Therefore,the pruning aims of Bing are to improve the proportion of bouquet branches by appropriate control measures which are significantly lighter than those in Longguan on the degree of control,to control the development of long fruit branches,and to improve the fruit setting rate.On the basis of thorough understand over characteristics of different sweet cherry varieties,targeted management measures can be developed.

Table 1 Survey results of fruit branch composition,fruit setting rates and yields of different sweet cherry varieties

Table 2 T-test results of branch type composition,fruit setting rate and yield of different sweet cherry variety

Cluster analysis is an effective toolfordealing with multi-variable sample classification.It has been extended from the field of biometrics to the main element of soft computing.In addition to hierarchical clustering and dynamic clustering,the currently established and applied clustering methods include self-organizing competitive neural network-based clustering[12],fuzzy clustering[13]and grey clustering[14-15].Nieet al.[16]overviewed the application of cluster analysis in researches on Chinese fruit,and they put forward opinions on indictors and classification number in cluster computing.Zhouet al.[17]divided the 8 wild cherryplum cultivars into 3 large groups by cluster analysis and other methods.The cluster results agreed with the actual situation of geographical distribution, and were also matched with the environmental features and habitat of distribution area.Huanget al.[18]conducted a cluster analysis for the fruit characteristics of 130 longan samples.The 130 test materials were divided into 5 groups,and there were great differences among the 5 groups.According to the appearance of 16 described traits and 11 measured traits,Liuet al.[19]clustered the 37 seedless grape varieties into 3 large groups and 5 sub groups.Wanget al.[20]determined the phenolics contents in major cherry cultivars in Shanxi Province,and they proposed quantitative indicators for flavor quality grading of cherry fruit.The results above indicate that cluster analysis is an effective tool for classification,but the efficiency of calculation result is different among different calculation methods(including conversion of data,definition of distance and clustering method).If DPS was adopted as the clustering tool and the original data are given,total 216(4×6×9)kinds of calculation results will be obtained for cluster analysis.Obviously,the comprehensive analysis of calculation results and professional features contributes to drawing of reliable conclusions.The selection and determination of variable type and quantity is of great significance for calculation results of cluster analysis.Nieet al.[16]considered that under the premise of effective classification of objectives,the traits should be as fewer as possible,while the clustering should be in line with the actual situation and has application value.In conclusion,the design of reasonable calculation program is a premise of obtaining effective and reasonable calculation results in cluster analysis about fruit;if the original variables are linearly independent on each other,the calculation results will be more desirable;in the case of a large number of variables,clustering can be conducted on the basis of appropriate dimension reduction,and the number of variables should be better less than 20;the tree map should be characterized by hierarchical tree structure;for clustering results,the indicators,belonging to differenttypes,should meet the requirement of"clear attribution,significantdifference".In this study,among the 8 indicators,5 indicators show significantdifferences among the 14 test materials by T-tests,which is in line with the expected research result.

[1]YANG J(杨军),SUN Y(孙怡).Comprehensive judgment on economic characters of Chinese cherry varieties by theory of grey system(中国樱桃品种经济性状的综合评判)[J].Journal of Biomathematics(生物数学学报),1998,13(3):334-337.

[2]YAN GH(闫国华),ZHANG KC(张开春),ZHOU Y (周宇),et al.Advances in research on health benefits of cherry(樱桃保健功能研究进展)[J].Journal of Food Science and Technology(食品工业科技),2008,29(2):313-316.

[3]AN CX(艾呈祥),XIN L(辛力),YU XM(余贤美),et al.Analysis of genetic diversity in Cerasus by SSR markers(樱桃主栽品种的遗传多样性分析)[J].Acta Horticulturae Sinica(园艺学报),2007,34(4):871-876.

[4]GAO J(高佳),WANG BG(王宝刚),FENG XY(冯晓元),et al.Composite appreciation of fruit characters in sweet cherry and sour cherry cultivars(甜樱桃和酸樱桃品种果实性状的综合评价)[J].Northern Horticulture(北方园艺),2011,17:17-21.

[5]LI WS(李文生),ZHANG KC(张开春),ZHANG XM(张晓明),et al.Botanical characteristics and identification of sweet cherry varieties(甜樱桃品种的植物学性状与品种鉴定)[J].Deciduous Fruits(落叶果树),2005,2:15-16.

[6]GU YQ(谷玉强),ZHANG ZM(张中民),ZHANG YX(张英霞),et al.Pruning technology of heart-shaped tree type with three major branches stratified for sweet cherry(甜樱桃三主枝分层开心形的整形修剪技术)[J].Northern Fruits(北方果树),2009,3:32-33.

[7]WEI HR(魏海蓉),LIU QZ(刘庆忠),ZHANG LS(张力思),et al.Pruning technology of small and sparse canopy tree type for dwarfing sweet cherry (矮砧甜樱桃小冠疏层形整形修剪技术)[J].Deciduous Fruits(落叶果树),2008,4:13-14.

[8]LIU MY(刘美英),ZHANG FX(张福兴),SUN QT(孙庆田),et al.Standardized slender spindle tree structure and pruning techniques of sweet cherry(甜樱桃细长纺锤形标准化树体结构与修剪技术)[J].China Fruits(中国果树),2013,2:48-49.

[9]ZHANG FX(张福兴),LIU MY(刘美英),SUN QT (孙庆田).Standardized freedom spindle tree structure and pruning techniques of sweet cherry(自由纺锤形树的甜樱桃标准树体结构及整形修剪技术)[J].Deciduous Fruits(落叶果树),2013,45(2):30-31.

[10]WEI GQ(魏国芹),SUN YG(孙玉刚),AN M(安淼),et al.Analysis on growth and bearing habits in different sweet cherry cultivars(甜樱桃不同品种生长结果习性调查)[J].Northern Horticulture(北方园艺),2012,6:13-15.

[11]YANG XH(杨晓华),DAO GL(戴桂林),NIE GW(聂国伟),et al.The effects of shoots constitute on the yield of sweet cherry(甜樱桃枝类组成及其对产量形成的影响)[J].Journal of Shanxi Agricultural sciences(山西农业科学),2014,42(6):568-572.

[12]CHENG L(程磊),LI SW(李绍稳),FAN JF (樊建峰),et al.The cluster analysis of red-seed watermelon inbred line based on self-organizing competitive neural networks(基于自组织竞争神经网络的红籽瓜自交系聚类分析)[J].Agriculture Network Information(农业网络信息),2008,5:27-31.

[13]LIU L(刘莉),LI SW(李绍稳).A Dangshan Su Pear scab forecasting expert system based on fuzzy clustering (模糊聚类在砀山酥梨黑星病预测专家系统中的应用)[J].Agriculture Network Information(农业网络信息),2004,2:12-14.

[14]LIU GP(刘桂苹),JIA YQ(贾永全),WANG CY(王长远).Application of grey cluster in husbandry production regionalization in Heilongjiang Province(灰色聚类在黑龙江省畜牧生产区划上的应用)[J].Heilongjiang Animal Science and Veterinary Medicine (黑龙江畜牧兽医),2003,9:25-27.

[15]SUN HY(孙浩元),XU JR(续九如),WANG YZ(王玉柱),et al.Selection of suitable climatic region for introduction of Ziziphus mauritiana Lam.using the methods of gray clustering and gray connecting analysis(应用灰色聚类和灰色关联分析对台湾青枣引种适宜区的选择)[J].Chinese Agricultural Science Bulletin(中国农学通报),2006,22(4):143-147.

[16]NIE JP(聂继云),ZHANG HJ(张红军),MA ZY(马智勇),et al.The application of cluster analysis in the fruit research in China and its problems(聚类分析在我国果树研究中的应用及问题分析)[J].Journal of Fruit Science(果树学报),2000,17(2):128-130.

[17]ZHOU L(周龙).Study on the communities characteristics and the genetic diversity of germplasm resource for Prunus divaricate(野生樱桃李群落特征及种质资源遗传多样性研究)[D].U-rumqi:Xinjiang Agricultural University(乌鲁木齐:新疆农业大学),2011.

[18]HUANG AP(黄爱萍),CHEN XP(陈秀萍),HU WS(胡文舜),et al.Fruit character diversity analysis and numerical classification of longan(DimocarpusLour.)germplasm resources(龙眼种质资源果实性状多样性分析及其数量分类研究)[J].Journal of Fruit Science(果树学报),2010,27(6):938-945.

[19]LIU CH(刘崇怀),TIE F(帖锋),PAN X(潘兴),et al.Cluster analysis of seedless grape cultivars(无核葡萄品种资源性状的聚类分析)[J].Journal of Fruit Science(果树学报),2006,23(4):531-533.

[20]WANG XP(王贤萍),DUAN ZM(段泽敏),DAI GL(戴桂林),et al.Polyphenol quantitative analysis and quality evaluated in fruit of sweet cherry cultivars(甜樱桃主要栽培品种多酚含量的测定与品质分析)[J].Chinese Agricultural Science Bulletin (中国农学通报),2011,27(13):173-176.