Procedures for Standard Evaluation and Data Management of Advanced Potato Clones Module 2. Healthy Tuber Yield Trials International Cooperators’ Guide

2|P á g i n a

© International Potato Center (CIP), 2014 ISBN: 978-92-9060-448-8 DOI: 10.4160/9789290604488 Digital version CIP publications contribute to important development information to the public arena. Readers are encouraged to quote or reproduce material from them in their own publications. As copyright holder, CIP requests acknowledgement and a copy of the publication where the citation or material appears. Please send a copy to the Communication and Public Awareness Department at the address below. International Potato Center P.O. Box 1558, Lima 12, Peru cip@cgiar.org • www.cipotato.org Citation De Haan, S.; Forbes, A.; Amoros, W.; Gastelo M.; Salas, E.; Hualla V.; De Mendiburu F.; Bonierbale M. 2014. Procedures for Standard Evaluation and Data Management of Advanced Potato Clones. Module 2: Healthy Tuber Yield Trials. International Cooperators Guide. Lima (Peru). International Potato Center. 44 p. Edition and Layout Sofia Tejada Setiembre 2014

3|P á g i n a

INDEX

INTRODUCTION 5 MODULE 2: HEALTHY TUBER YIELD TRIALS 7 LOCATION 7 GENETIC MATERIALS 7 EXPERIMENTAL DESIGN 8 OTHERS DESIGNS 9 FIELD MANAGEMENT AND INFORMATION ON ENVIRONMENTAL FACTORS 11 GEOGRAPHICAL INFORMATION SYSTEM (GIS) – TAKING WAYPOINT 11 EVALUATION PARAMETERS 11 Period of vegetative Development 11 Number of tubers planted (NTP) 11 Number of emerged plants/Plot (NPE) 11 Plant growth Habit (PGH) 11 Plant uniformity (Plant_Unif) 13 Plant vigor (Plant_Vigor) 14 Flowering degree (Flowering) 15 Senescence (SE) 16 Period Of Harvest 17 Number of plants harvested (NPH) 17 Number of stolons (Num_Stolon) 17 Lenght of stolons (Leng_Stolon) 18 Tuber appearance (Tuber_Apper) 19 Tuber uniformity (Tub_Unif) 20 Tuber Size (Tub_Size) 21 Number marketable tubers category I/plot (NMTCI) 22 Number marketable tubers category ii/plot (NMTCII) 22 Number of non-marketable tubers/plot (NNOMTP) 22 Marketable tuber weight category I/plot (MTWCI) 22 Marketable tuber weight category II/plot (MTWCII) 22 Non-marketable tuber weight/plot (NOMTWP) 22 CALCULATED OF VARIABLES 23 OTHER EVALUATIONS 24 EXTERNAL DEFECTS 24 INTERNAL PROBLEMS 25 PERCENTAGE OF DEFECTED TUBERS 25 DATA RECORDING 25 DATA ANALYSIS 26 QUANTITATIVE CONTINUOUS/DISCRETE VARIABLES 26 QUANTITATIVE ORDINAL (PSEUDO- QUANTITATIVE) VARIABLE 26 QUALITATIVE ORDINAL VARIABLES 26 EXAMPLE AND DATA INTERPRETATION 27 VALIDATION OF THE EXPERIMENT 27 SELECTION CRITERIA 27 FILLING THE DATACOLLECTOR TEMPLATE OF HEALTHY TUBER YIELD 29 FORM: MINIMAL 29 FORM: INSTALLATION 30

4|P á g i n a

FORM: MATERIAL_LIST 31 FORM: CROP MANAGEMENT 31 FORMS: HOBO DATA, WEATHER DATA, SOIL ANALYSIS 32 FORM: VAR_LIST 33 FORM: FIELDBOOK 34 ANNEX 1 35 SPLIT-PLOT DESIGN 35 Characteristics 35 Randomization 35 PARTIALLY BALANCED LATTICE DESIGN 36 Characteristics 36 Randomization 36 EXPERIMENTAL DESIGN FOR LARGE NUMBER OF GENOTYPES 38 ALPHA-LATTICE 39 Characteristics 39 Randomization 39 AUGMENTED BLOCK DESIGN 41 Characteristics 41 Randomization 41 BIBLIOGRAPHY 43

5|P á g i n a

I.- INTRODUCTION

The International Potato Center (CIP) currently manages potato selection trials in no less than 50 different countries in Africa, Asia and Latin America. Each region or sub-region has scientists in charge of monitoring breeding advances and varietal selection. The procedures presented in this basic guide are designed to assist CIP staff in organizing trials and data collection in such a way that data can be shared, centrally stored and uploaded to the Global Trial Data Management System. Among breeders and collaborators we do at least need agreements about: (i) the most important traits to be observed and measured, (ii) standardized procedures and formats to record data, (iii) a user-friendly and practical global system to upload, store and share data.

6|P á g i n a

7|P á g i n a

II.- MODULE 2: HEALTHY TUBER YIELD TRIALS

Healthy Tuber Yield Trials can be conducted with 1 up to 30 clones and is recommended for advanced materials that have already shown superior performance in intentional exposure trials for key traits. Location

Genetic Materials Clones or varieties from CIP and/or national breeding programs can be evaluated. At least two of the most commonly-used varieties should be used as controls. High-quality seed of the same origin should be used as control varieties and clones. The tuber yield trial in the first season requires at least 40 seed tubers per entry (10 plants per row), to be planted in three replications in one location. During the following seasons, the plot size and number of locations should be increased depending on seed availability.

Season 1: During the first season, yield trials are established in a location representative of the targeted production area. However, the number and quality of the seed used might force this first evaluation to be located in an experimental station.

Season 2 onward: Yield trials are established in one or more locations representative of the targeted production areas. Yield trials can be combined with on and off-farm experiments, management trials, participatory selection and/or GxE interaction trials.

8|P á g i n a

Experimental Design The Healthy Tuber Yield Trials uses a Randomized Complete Block Design (RCBD), where replications of clones are planted in blocks and within each block all genotypes are randomized. When the pattern of field variability is unidirectional, long and narrow blocks should be used. When the pattern of variability is not predictable, blocks should be as square or rectangular plots of double or multiple rows. These are preferable to long, single row plots. Single row plot should not be carried out because of the inter-plot competition (border affects due to neighbor plot within a block). Ideally, Healthy Tuber Yield Trials must be carried out in at least three locations. The clear advantage to conduct tuber yield trials in three locations is that this saves time, because in potato trials temporal variation of test environments can be replaced by spatial variation of test environments (locations). Trials carried out across locations allow for the separation of effects due to genotypes, genotype by environment interaction and plot error. Furthermore, with three divergent locations it is possible to determine for each genotype stability parameters, which must be considered as an additional character associated with yield. The randomization process for a RCBD design is applied to each of the blocks. Randomizing can be done with “DataCollector”.

In the RCBD design, all treatments (advanced clones/varieties) are grouped into uniform blocks of equal size. The main purpose of blocking is to reduce experimental error by eliminating sources of heterogeneity such as soil fertility or field slopes. With a predictable pattern of field variability, plot shape and block orientation can be carefully chosen so that the experimental conditions within each block are as uniform as possible.

9|P á g i n a

Analysis of Variance (ANOVA) is used to analyze the data collected in a RCBD. The three sources of variability used in the statistical model are the treatment (variety/potato clone), the blocks (repetition) and the experimental error for each environment.

Others Designs Depending on the type of trial and its objectives, additional trial designs (such as a split-plot design and lattice design) can be applied. Brief descriptions and the randomization processes of those designs are provided in the annex1. Further information can also be obtained from technical manuals dealing with experimental trial designs (Gomez and Gomez, 1984).

10|P á g i n a

11|P á g i n a

IRONMENTAL FACTORS III.- FIELD MANAGEMENT AND INFORMATION ON ENVIRONMENTAL FACTORS

Field management should follow standard agronomic practices and local procedures to protect the crop from pests and diseases. Meteorological data and soil analyses are ideally collected as to identify spatial patterns among experimental sides and agro-ecological zones. Climatic data may be easily accessible only for on-station trials, whereas availability for other experiments may depend on the proximity of the test site to a meteorological station. Geographical Information System (GIS) – taking waypoint For plant breeders, the strength of spatial data management systems is its capacity to provide information on test location that can be used in supporting the analysis of genotype x environment interactions. Ideally a so-called waypoint is taken with a GIS device to record longitude, latitude and altitude for each trials site. Evaluation parameters Once the Healthy Tuber Yield Trial(s) have been established, the following agronomic data should be collected during:

Period of vegetative development a) Number of tubers planted (NTP): This is recorded directly at planting. b) Number of emerged plants/plot (NPE): This evaluation is performed 45 days after planting, count the number of emerged plants by plot.

12|P á g i n a

c) Plant growth habit (PGH)1: This evaluation is performed 45 days after planting using a scale 1 to 3. (Gomez, 2004).

1 It is not necessary to collect this variable for multiple plots or years. Once the plant habit is established no re-recording is required.

Scale State Description

1 Erect The stems are almost vertical and the angle of insertion between the leaves rachis and the main stem is sharp, around 30°.

2 Semi-erect

The stems have more or less a vertical growth, but some secondary stems open up a bit and the insertion angle between the leaves rachis and the main stem is more open, around 45°.

3 Decumbent

The stems are more open, some secondary stems are open to the point of reaching the ground. From there the stems tend to recover some vertical growth. The angle of insertion of the leaf rachis with the main stem is very open, from 60 to 90 °. Such plants cover the ground very well and have most of the leaf area exposed to sunlight.

Erect Semi-erect Decumbent 1 2 3

13|P á g i n a

Very heterogeneous Heterogeneous Intermediate

Uniform Very uniform

3 5 1

7 9

d) Plant uniformity (Plant_Unif): This evalutation is performed 45 days after planting and should be evaluated using a scale from 1 to 9. (Salas et al., 2004)

Scale State Description

1 Very heterogeneous Height, vigor, growth stage very heterogeneous.

3 Heterogeneous 75% of the plants show height, vigor and growth stage heterogeneity.

5 Intermediate 50% of the plants show height, vigor and growth stage heterogeneity.

7 Uniform 75% of the plants show height, vigor and growth stage homogeneity.

9 Very uniform 100% of the plants show height, vigor, growth stage homogeneity.

14|P á g i n a

Very weak Weak

Medium Vigorous Very vigorous

3

5

1

7 9

e) Plant vigor (Plant_Vigor): This evaluation is performed 45 days after planting and should be evaluated using a scale from 1 to 9. (Salas et al., 2004).

Scale State Description

1 Very weak All the plants are small (< 20 cm), few leaves, weak plants, very thin stems and/or light green color.

3 Weak 75% of the plants are small (< 20 cm) or all the plants are between 20 and 30 cm, the plants have few leaves, thin stems and/or light green color.

5 Medium Intermediate or normal.

7 Vigorous 75% of the plants are over 50 cm, robust with foliage of dark green color, thick stems and leaves very well developed.

9 Very vigorous

All the plants are over 70 cm and ground coverage is complete. The plants are robust, with thick stems and abundant foliage of dark green color

15|P á g i n a

No Bud Aborted Bud Low

Moderate Profuse

1 3 0

5 7

f) Flowering degree (Flowering)2: This evaluation is preformed 60 days after planting and recorded using a scale from 1 to 7 (Bioversity & CIP, 2009; Gomez, 2004)

2 It is not necessary to collect this variable for multiple plots or years. Once the flowering degree is established no re-recording is required.

Scale State Description

0 No bud No inflorescence although inflorescence are rudimentary and consequently of buttons.

1 Aborted bud Presence of small or rudimentary inflorescences that can show an abortion or abscission point at the joint of the pedicel.

3 Low Flowering is scarce with the presence of 2 to 3 flowers (buds, flower buds, flowers, fruits and flower abscissions) per inflorescence.

5 Moderate Flowering is moderate with 8 to 12 flowers (buds, flower buds, flowers, fruits and flower abscissions) per inflorescence.

7 Profuse Profuse flowering with 20 or more flowers (buds, flower buds, flowers, fruits and flower abscissions) per inflorescence.

16|P á g i n a

Very late Late

Early Very early

3 5 1

7 9

Medium

g) Senescence (SE): This evaluation is performed 70 or 90 days after planting (Depending on the vegetative period of the genotypes) and should be evaluated using a scale from 1 to 9. (Amoros & Gastelo, 2011. Personal communication)

Scale State Description

1 Very late All the plants still show green foliage and flowers

3 Late Most of the plants are still green, flowering is over and berries might be formed.

5 Medium

The plants are still being green or on the onset of senescence, there may be a slight yellowing. The angle of insertion of the leaves on the stems may have become more obtuse than in the younger plants of the same clone. The formation of berries can be advanced and abundant in fertile clones.

7 Early

The plants have senescent foliage, yellowing is more advanced but the stems may still be upright. If berries are present, their color will turn from green to pale green or yellow green.

9 Very early The plants are completely senescent, yellowing is complete and uniform, and the stems are decumbent.

17|P á g i n a

Certainly observations on disease and pest damage can also be recorded. Because the purpose of the trial is to evaluate yield under standard or optimum crop management, integrated crop management practices should be used to control pests and diseases.

The harvest period The foliage should be burned or cut 10 to 15 days prior to harvesting. It is recommended that evaluations are conducted in the following sequence: a) Number of plants harvested (NPH) b) Number of stolons (Num_Stolon)3: Overall assessment of the number and length of the stolons based on inspection of the stolons using a 1 to 9 scale. (Amoros & Gastelo, 2011. Personal communication)

Scale State Description

1 Very few Plants show no stolon or very few (0 to 4).

3 Few Plants with 5 to 10 stolons.

5 Medium Plants with 11 to 15 stolons.

7 High Plants with 16 to 25 stolons.

9 Very high Plants with more than 25 stolons.

3 It is not necessary to collect this variable for multiple plots or years. Once the variable is established no re-recording is required.

18|P á g i n a

Very short Short

Long Very long

Medium

3 5 1

9 7

c) Lenght of stolons (Leng_Stolon)4

Scale State Description

1 Very short X ≤ 20 cm long.

3 Short 20 cm <X ≤ 40 cm long.

5 Medium 40 cm < X ≤ 60 cm long.

7 Long 60 cm < X ≤ 80 cm long.

9 Very long X > 80 cm long.

4 It is not necessary to collect this variable for multiple plots or years. Once the variable is established no re-recording is required.

19|P á g i n a

Very poor Very good Regular 9 5 1

d) Tuber Appearance (Tuber_Apper): (Amoros & Gastelo, 2011. Personal communication) Scale State Description

1 Very poor Very low yield, totally misshapes and non-uniform size

3 Poor Low yield, some misshapes but non uniform size

5 Regular Medium yield, good shape but non uniform size

7 Good Good yield, good shape and uniform size

9 Very good High yield, good shape and very uniform size

20|P á g i n a

Very heterogeneous Very uniform Intermediate 9 5 1

e) Tuber uniformity (Tub_Unif): Overall assessment of tuber uniformity is based on the inspection of the harvested tubers using a 1 to 9 scale. (Amoros & Gastelo, 2011. Personal communication)

Scale State Description

1 Very heterogeneous All tuber sizes are present (from very small to large)

3 Heterogeneous All tuber sizes are present but there is a predominant size

5 Intermediate There are only 2 or 3 tuber sizes with a predominant size

7 Uniform Only two sizes are present with a predominant tuber size

9 Very uniform Only one tuber size

21|P á g i n a

Very small Small

Large Very large

Medium 3 5 1

9 7

f) Tuber size (Tub_size): Data is collected using a 1 to 9 scale. (Amoros & Gastelo, 2011. Personal communication)

Scale State Description

1 Very small Most tubers are very small (<2cm).

3 Small Tubers are small, between 2 and 4cm.

5 Medium Tubers are between 4 and 6cm

7 Large Tubers are large, between 6 and 9 cm.

9 Very large Tubers are over 9 cm.

22|P á g i n a

g) Number marketable tubers category I/plot (NMTCI): Count the number of marketable tubers for category I with weighing between 200-300 g or tubers of 60 mm. h) Number marketable tubers category II/plot (NMTCII): Count the number of marketable tubers category II with weighing between 80-200 g or tubers between 30-60 mm. These categories I and II are arbitrary and can be change according to the country or region where are being evaluated. Each evaluator is free to use locally relevant criteria; however, each category should be defined in order to facilitate comparison of data between countries. i) Number of non-marketable tubers/plot (NNoMTP): Count the number of non marketable tubers with weighing less of 80 g or less of 30 mm. j) Marketable tuber weight category I/plot (MTWCI): Weigh marketable tuber category I/plot. The unit of measure is Kilograms. k)Marketable tuber weight category II/plot (MTWCII): Weigh the marketable tuber category II/plot. The unit of measure is Kilograms. m) Non-marketable tuber weight/plot (NoMTWP): Weigh the non-marketable tuber/plot. The unit of measure is Kilograms. Data should be checked for any errors made during collection or transcription and checked as soon as possible afterwards so that corrections can be made where necessary.

23|P á g i n a

Calculated of variables Several variables can be derived from the raw data the Healthy Tuber Yield Trials. We consider: Total Tuber Yield, Marketable Tuber Yield and Average Tuber Weight as a must.

Variable Abbrevia-

tions Unit Formula

Percentage of Plants Emerged

PPE Percentage 𝐏𝐏𝐄 =NPE ∗ 100

NTP

Percentage of Plants Harvested

PPH Percentage 𝐏𝐏𝐇 =NPH ∗ 100

NTP

Number Marketable Tubers/Plot

NMTP Count 𝐍𝐌𝐓𝐏 = NMTCI + NMTCII

Total Number of Tubers/Plot

TNTP Count 𝐓𝐍𝐓𝐏 = NMTP + NNoMTP

Total Number of Tubers/Plant

TNTPL Count 𝐓𝐍𝐓𝐏𝐋 =TNTPNPH

Number Marketable Tubers/Plant

NMTPL Count 𝐍𝐌𝐓𝐏𝐋 =NMTPNPH

Total Tuber Weight/Plot

TTWP kg 𝐓𝐓𝐖𝐏 = MTWCI + MTWCII

+ NoMTWP

Total Tuber Weight/Plant

TTWPL kg/pl 𝐓𝐓𝐖𝐏𝐋 =TTWPNPH

Total Tuber Yield Adjusted

TTYA tons/ha 𝐓𝐓𝐘𝐀 =TTWPL ∗ PLD

1000

Total Tuber Yield No Adjusted

TTYNA tons/ha 𝐓𝐓𝐘𝐍𝐀 = �TTWP

PLS� ∗ 10

24|P á g i n a

Variable Abbrevia-

tions Unit Formula

Marketable Tuber Weight/Plot

MTWP kg 𝐌𝐓𝐖𝐏 = MTWCI + MTWCII

Marketable Tuber Weight/Plant

MTWPL kg/pl 𝐌𝐓𝐖𝐏𝐋 =MTWP

NPH

Marketable Tuber Yield Adjusted

MTYA tons/ha 𝐌𝐓𝐘𝐀 =MTWPL ∗ PLD

1000

Marketable Tuber Yield No Adjusted

MTYNA tons/ha 𝐌𝐓𝐘𝐍𝐀 = �MTWP

PLS� ∗ 10

Average Tuber Weight

ATW g 𝐀𝐓𝐖 = �TTWPTNTP

� ∗ 1000

Average marketable tuber weight

ATMW g 𝐀𝐓𝐌𝐖 = �MTWPNMTP

� ∗ 1000

Where: PLS= Net plot size and PLD=Planting Density.

Other evaluations A random sample of 10 tubers per clone should be cut transversally and checked for: a) External defects: such as cracking, secondary growth and warts, and

25|P á g i n a

b) Internal problems: Such as hollow heart, black spots, heat necrosis, and rot. Internal defects should be reported at harvest time. This is critical for estimating processing quality. c) Percentage of defected tubers: For each entry, the number of affected tubers is recorded on the tuber yield datasheet. Data recording

CIP together with partners are promoting the use of Data Collector

software that helps to standardize and ensure data quality (Simon

et al., 2012); it is part of the International Potato Center’s Global

Data Management System and assists researchers in data analysis

by automatically calculating the variables for Healthy Tuber Yield

Trials. The information should be recorded onto form

DataCollector.

Phase Component Method Registration Form

Field Installation and characterization to define selection environments

Minimal - basic data List Minimal

Experimental design List Installation

Field management and evaluation dates

List Crop_Management

List of materials List Material_list

Climate data Weather station Weather_data

Soil analysis Soil analysis Soil_Analysis

Experiment results

Observed and calculated variables Fieldbook

26|P á g i n a

Data analysis Types of variables: Data Quality Control: Simple statistics such as mean, standard error, frequency distribution and boxplots should be used to explore the data. Yield data are analyzed using variance analysis (ANOVA) and means are compared using statistical comparison tests such as LSD, Tukey, Waller-Duncan, and Bonferroni. Orthogonal contrasts and Dunnett tests can be used to compare the advanced clones with the control(s). The analysis of residuals is recommended to test the validity of the model and to analyze the behavior of the variance (homogeneous or not). All analysis can be performed using

Quantitative continuous and discrete variables: Numeric variables following approximately a normal distribution (e.g. Total Tuber Yield, Dry Matter, etc...) are analyzed using parametric statistics.

Quantitative ordinal (pseudo- quantitative) variable: Numeric variables which show in their distribution strong deviation from a normal distribution (e.g. The percent of plant infection (which is used in evaluating clonal resistance to a disease). This variable, which represents the evaluator’s estimation of the damage, is more a rank than a measurement.

Qualitative ordinal variables: Data cannot be measured, they are ranked or attached to a rating scale. (e.g. Scores with a scale of 1 to 9 for plant uniformity or scores with a scale 1 to 3 for plant growth habit). Ordinal variables are analyzed and compared using non-parametric methods of analysis.

27|P á g i n a

DataCollector Software or other statistical packages that facilitates analysis and reports of the results. Example and Data interpretation Validation of the experiment: An experimental trial for tuber yield evaluation is considered to have been carried out under appropriate conditions if the experiment’s coefficient of variation does not exceed 30%. Selection criteria: Performance of each advanced clone is compared with the performance of the control(s). It is important to consider the commercial yield of the entry rather than the total yield. In most situations, the ability of a clone to develop numerous small tubers will be viewed as a negative characteristic.

28|P á g i n a

29|P á g i n a

IV.- FILLING THE DataCollector TEMPLATE OF HEALTHY TUBER YIELD

Form: Minimal DataCollector software will complete this information according with your locality. Be sure to complete the “Begin date” and the “End date”. Please write and single quote before the date The correct format date is: ‘yyyy-mm-dd. e.g. ‘2014-04-07

[Back to “Data Recording” table]

30|P á g i n a

Form: Installation Please complete this form according your experimental design. Note: Consider the net plot size.

[Back to “Data Recording” table]

31|P á g i n a

Form: Material_List Here you need to complete the code clones in the column "Institutional number", pedigree information and mark with "x" the clone or genotype control.

Form: Crop management Summarize all procedures that were performer in the experiment (date of planting, evaluation dates and all field management data that you consider important. The correct format date is: yyyy-mm-dd, please write and single quote before the date. e.g. ‘2014-04-07.

[Back to “Data Recording” table]

32|P á g i n a

Forms: Soil analysis & Weather data Complete this information with your experiemtnal data.

[Back to “Data Recording” table]

33|P á g i n a

Form: Var_List Type on Selection direction (+) and on Selection weight (1) depending on the variable to analyze.

[Back to “Data Recording” table]

34|P á g i n a

Form: Fieldbook Allows the entry of observed and calculated data of the variables that were performed according to the experimental design.

[Back to “Data Recording” table]

35|P á g i n a

V.- ANNEX 1 A brief description of commonly used incomplete block designs

Split-Plot Design Characteristics The split-plot design is a special kind of incomplete block design. The underlying principle of the split-plot design is that whole plots, subject to one or more treatments (factor A), are divided into subplots to which one or more additional treatments are applied (factor B). Thus, each whole plot may be considered as a block for subplot treatments (factor B), but only as an incomplete block as far as the full set of treatments is concerned (factor A + B). The design may be used when an additional factor (such as planting density or fertilizer use) is to be incorporated into an experiment to increase its scope. Randomization Randomization is a two-stage process. First, factor A treatments are randomized over the whole plot; then factor B treatments are randomized within the subplots.

36|P á g i n a

Partially Balanced Lattice Design Characteristics The partially balanced lattice design is recommended when the number of treatments is very large or when the experimental units are very heterogeneous. Lattice designs are incomplete block designs. Each block does not contain all treatments, so the precision of comparison between treatments differs depending if the treatments belong to the same block or not. The lattice design (also called double lattice or square lattice), is a partially balanced design in which the number of treatments is a perfect square (9, 16, 25, 36, 49, 64, 81, 121 etc.) and the number of treatments within each block is equal to the square-root of the total number of treatments. This design needs two or multiples of two replications. The experimental units within each incomplete block should be as homogeneous as possible. Randomizing Treatments are arranged in the form of a square (step 1). Treatments are grouped by row, and then by columns. The row grouping is generally known as X grouping. The group of treatments in one row will form a block.

All the rows (blocks) will make one repetition (step 2). The column grouping is generally known as Y grouping. The group of treatments in one column will constitute another block. This Y grouping will form the other repetition (step 3). The X grouping

37|P á g i n a

and Y grouping ensure that treatments occurring together in the same block once do not appear together in the same block again.

For each repetition, the randomization is a three-stage process: the blocks are randomized, each treatment is randomized within each block (step 4), and ultimately a treatment is randomly assigned to each plot.

38|P á g i n a

Experimental design for large number of genotypes

Phenotypic data are vitally important for assessment of the within-environment error structure for each of the trials that will be used later in the MET analysis. Therefore, appropriate control of local variability through efficient experimental design is of key importance. Spatial variability in the field is a universal phenomenon that affects the detection of differences among treatments in agricultural experiments by inflating the estimated experimental error variance. Recently, efficient experimental designs (both unreplicated and replicated) have been developed, assuming that observations are not independent in that contiguous plots in the field may be spatially correlated (Cullis et al., 2006; Martin et al., 2004).

39|P á g i n a

Alpha-lattice Characteristics It is common in many experiments have a large number of seedlots and a small number of replicates. In this case, one of the most suitable design is the Alpha-lattice; an incomplete block design that divide the replicates into incomplete blocks that contain a fraction of the total number of treatments. Treatments are distributed among the blocks so that all pairs occur in the same incomplete-block in nearly equal frequency. The breeder should beer in mind these conditions (Patterson & Williams, 1976):

→ The number of genotypes (v) must be multiple of the size

block(k).

→ The number of block (s) is calculated dividing the number of

genotypes and the block size: s=v/k .

→ The parameters for Alpha design are divided in four cases: Case I: When r=2 and k ≤ s; Case II: When r=3, the number of blocks(s) is odd and

k ≤ s. Case III: When r=3, the number of blocks (s) IS even and

k ≤ s-1; Case IV: When r=4, the number of blocks (s) is odd but

not a multiple of 3, and k ≤ s.

→ The range of number of replication (r) can vary from two

until four (r=2,3,4).

40|P á g i n a

Randomizing

→ Example 1 An example of the randomization in Alpha design with two replication (r = 2), 9 treatments (v = 9) and block size (k = 3). Thus, the number of blocks is s =3.

→ Example 2

Another example with T=32 treatments {T1, T2, T3,…,T32}, two replications (r=2) and block size k=4. Thus the number of block is s=8.

A graphical representation of the example 2,

41|P á g i n a

Augmented block design Characteristics

The augmented designs are very useful for testing many genotypes or clones but taking as limiting the number of seed tubers. These designs just make one repetition for treatment and include checks or controls which are systematically replicated in each block to control the heterogeneity of the environment. The repeated checks measure the spatial variation and the experimental units (without repetition) are assessed on the basis of adjacent checks. The experiments are usually conducted using an augmented randomized complete block design, even more incomplete block designs. Randomizing Let c: number of different checks per block Let r: number of blocks=number of replicates of a check Let V = 30 genotipos {V1,V2,…..V30} and the number of checks(#checks) = 3 {a,b,c} Number of blocks: ((10)/(c-1))+1 = (10/2)+1 = 6

42|P á g i n a

43|P á g i n a

VI.- BIBLIOGRAPHY Bioversity International; International Potato Center (CIP) 2009. Key access and utilization descriptors for cultivated potato genetic resources. http://www.bioversityinternational.org/nc/publications/publication/issue/key_access_and_utilization_descriptors_for_cultivated_potato_genetic_resources.html

Cullis, B. R.; Smith, A. B.; &Coombes, N. E. (2006). On the design of early generation variety trials with correlated data. Journal of Agricultural, Biological, and Environmental Statistics, 11 (4), 381-393.

International Potato Center (CIP). 2006. Procedures for standard evaluation trials of advanced potato clones. An International Cooperators’ Guide.

Gomez, K.A.; Gomez, A.A. 1984. Statistical Procedures in Agricultural Research, New York, Chichester, etc.: Wiley, 2nd edition, paperback, pp. 680 Gomez, R. (CIP). 2004. Guía para las caracterizaciones morfológicas básicas en colecciones de papas nativas: Sexta aproximación. Conservación In Situ de Cultivos Nativos y sus Parientes Silvestres INIEA. Seminario - Taller Nacional de Caracterización In Situ. Lima (Peru). INIEA. pp. 44-71. Martin, R.; Eccleston, J.; & Chan, B. (2004) Efficient factorial experiments when the data are spatially correlated. Journal of Statistical Planning and Inference, 126(1), 377-395.

Patterson, H.D. & Williams, E. R. (1976). A new class of resolvable incomplete

block designs. Biome- trika, 63(1), 83-92. Salas, E.; Mihovilovich, E.; Amoros, W.; Bonierbale, M. (CIP). 2004. Estimation of variability for PLRV resistance in tropically adapted potatoes. American Journal of Potato Research. (USA). ISSN 1099-209X. 81(1):86

Simon, R., Salas, E., Eyzaguirre, R., Hualla, V., de Haan, S., & Bonierbale, M.

(2012). Desarrollo de un software estadístico para estandarización y chequeo

de calidad de datos del campo en raíces y tubérculos (GDET4RT) para el

mejoramiento de cultivos. In 12th International Conference, ALAP, Uberlandia

Brazil.

44|P á g i n a

z

The International Potato Center (known by its Spanish acronym CIP) is a research-for-development organization with a focus on potato, sweetpotato, and Andean roots and tubers. CIP is dedicated to delivering sustainable sczzience-based solutions to the pressing world issues of hunger, poverty, gender equity, climate change and the preservation of our Earth’s fragile biodiversity and natural resources. www.cipotato.org CIP is a member of CGIAR. CGIAR is a global agriculture research partnership for a food-secure future. Its science is carried out by the 15 research centers who are members of the CGIAR Consortium in collaboration with hundreds of partner organizations. www.cgiar.org International Potato Center • Av. La Molina 1895, La Molina • Apartado 1558 Lima 12, Perú