Selection Principles

Part of Seed Saving

Selection is the mechanism of plant breeding — the deliberate shaping of a crop population toward desirable traits over successive generations. Understanding how different selection methods work, when to apply them, and how to balance competing selection pressures allows a seed saver to systematically improve crops rather than simply maintaining them.

The Genetic Basis of Selection

Every plant trait has a genetic component and an environmental component. A plant in a favorable soil position may appear more productive than its genetics warrant, while a genetically superior plant in poor soil may appear mediocre. Selection works by identifying the genetic component of variation — which plants perform well because of what they are, not because of where they grew.

This is the central challenge of plant breeding: distinguishing genetic differences from environmental noise.

Heritability is the proportion of observed variation in a trait that is attributable to genetics. Traits with high heritability (seed color, presence/absence of disease symptoms, flower structure) respond quickly to selection. Traits with low heritability (yield, root size) are heavily influenced by environment and require many generations of selection or careful experimental designs to improve.

Trait	Typical Heritability	Generations to See Progress
Seed color	Very high	1–2
Disease resistance (qualitative)	High	1–3
Maturity date	Medium-high	3–5
Drought tolerance	Medium	5–8
Yield per plant	Low-medium	5–10+
Fruit size	Medium	4–7
Flavor	Medium-low	5–10+

Mass Selection

Mass selection is the simplest and most widely practiced selection method. It involves choosing the best individual plants from a population and mixing their seed together to form the seed lot for the next generation.

How Mass Selection Works

1. Grow the full population
2. Evaluate all plants at the appropriate growth stage(s)
3. Designate the top percentage of plants as seed parents (typically top 10–25%)
4. Harvest and mix their seed together
5. Plant the mixed seed the following season
6. Repeat

The resulting population gradually shifts toward the traits you selected for, as alleles associated with those traits increase in frequency over generations.

Effectiveness of Mass Selection

Mass selection works best for:

• Traits with high heritability (color, disease resistance, maturity)
• Cross-pollinating species (corn, brassicas, carrots) — more genetic recombination each generation means faster response
• Traits that are consistent across environments — traits that look the same regardless of where in the field the plant grows

Mass selection is less effective for:

• Low-heritability traits (yield) — environmental variation masks genetic differences
• Self-pollinating crops (tomatoes, beans) — less recombination slows population-level response

Selection Intensity

Selection intensity is the percentage of plants saved as parents. Saving fewer plants (higher selection intensity) gives faster progress per generation but risks narrowing genetic diversity.

Selection Intensity	Plants Saved	Genetic Diversity	Progress per Generation
Very high	Top 5%	Narrow risk	Fast
High	Top 10%	Moderate	Moderate-fast
Moderate	Top 25%	Good	Moderate
Low	Top 50%	Broad	Slow
Minimal	Top 75%+	Very broad	Very slow

For survival seed saving, moderate selection (top 20–30%) balances improvement pace with genetic diversity retention. Reserve aggressive selection (top 5–10%) for traits where you see a genuine crisis (widespread disease in a susceptible population).

Family Selection

Family selection maintains each seed parent’s offspring separately, evaluates families (rows or plots) as units, and selects the best families for the next generation. More powerful than mass selection because it reduces environmental noise.

How Family Selection Works

1. Select seed from N individual plants in generation 1; keep each plant’s seed separate (labelled families)
2. In generation 2, plant each family in its own row, all in similar positions (avoid edges of field)
3. Evaluate the performance of each row as a whole — which row is most uniform, most productive, most disease-resistant?
4. Select the 2–4 best-performing rows (families)
5. From within those rows, select the best individual plants
6. Save seed from selected plants within selected families for generation 3

Why Family Selection Is More Powerful

A single plant may be tall because it occupies a fertile soil patch, not because it is genetically tall. But a family (all offspring of that tall plant) grown across an entire row that is uniformly taller than adjacent rows — that represents genuine genetic potential.

Family selection requires more space (separate rows per family) and more record keeping than mass selection. In a small garden, it may not be practical. In a larger field setting, it is the preferred method for traits with moderate-to-low heritability.

Pedigree Selection

Pedigree selection extends family selection into a multi-generation tracking system, maintaining full records of which plants and families produced which offspring at every stage.

Used in formal plant breeding programs. At small scale, useful for maintaining exact knowledge of a variety’s recent history and for tracking progress on multiple traits simultaneously.

Requires detailed written records at every stage, which is a commitment but creates invaluable documentation for long-term variety management.

Recurrent Selection

Recurrent selection is a cyclical scheme for cross-pollinating crops where superior plants are mated with each other repeatedly over many cycles, concentrating favorable alleles while maintaining enough diversity for further improvement.

Particularly effective for corn and other obligate cross-pollinators. One common form:

1. Grow population and select superior plants (S0 generation)
2. Allow selected plants to cross-pollinate freely
3. Evaluate offspring (S1)
4. Select the best S1 plants
5. Repeat

Each cycle increases the frequency of favorable alleles while keeping the population large enough to avoid inbreeding depression.

Balancing Multiple Traits

Real crops need to be good at many things simultaneously. Selecting hard for one trait while ignoring others can produce a variety that excels in one area but fails practically.

Index Selection

Assign weights to multiple traits and evaluate plants by a combined score. Example for a bean variety:

Trait	Weight	Plant A Score	Plant B Score
Yield	40%	80	70
Disease resistance	35%	60	85
Seed quality	25%	90	75
Index Total		76	76.75

Plant B scores slightly higher on the index despite lower yield because its disease resistance advantage outweighs the yield deficit according to the assigned weights.

Weight assignment is a judgment call based on your priorities. In a survival context, disease resistance and climate tolerance should receive higher weights than yield and cosmetic quality.

Tandem Selection

Select for one trait at a time, rotating between traits in successive generations. Simpler than index selection but slower for all traits.

Year 1: Select for disease resistance Year 2: Select for yield (among disease-resistant survivors) Year 3: Select for maturity date Year 4: Return to disease resistance

This prevents any single trait from being abandoned but requires patience — each trait only advances every 3–4 years.

Threshold Selection (Independent Culling)

Set a minimum threshold for each trait. Any plant that fails to meet any threshold is culled, regardless of how well it scores on other traits.

Example thresholds:

• Yield: must be above the population median
• Disease: must show less than 20% leaf area affected
• Maturity: must ripen before first expected frost

A plant with exceptional yield but severe disease is still culled. This method prevents extremes — a variety cannot sacrifice one critical trait to gain on another.

Negative vs. Positive Selection

Positive selection: Choosing the best plants to save seed from. Active, intentional, requires presence throughout the season.

Negative selection (roguing): Removing the worst plants before they can pollinate. Passive benefit — you don’t have to perfectly identify the best plants, just reliably remove the obvious worst.

Both work together. In a small operation, negative selection is often more practical because:

• Off-types and diseased plants are easier to identify than the “best” plants
• Roguing before flowering prevents contamination by bad pollen
• The remaining population can then set seed more freely

Roguing Complements, Not Replaces, Positive Selection

Roguing the worst plants raises the average quality of the population but does not drive it toward a specific goal. Positive selection from the best individuals adds directional improvement. Use both.

Long-term Adaptation Strategy

Beyond improving specific traits, the goal of multi-generational seed saving is to develop varieties genuinely adapted to your specific location — your soil type, microclimate, rainfall patterns, pest pressure, and season length.

This local adaptation cannot be purchased — it must be developed by growing seed in place, selecting under real local conditions, and allowing the population to respond to actual stresses over many generations.

A locally adapted landrace variety, even if it appears inferior to a commercial variety on paper, will often outperform that commercial variety in your specific conditions over a 10-year horizon, because it has been shaped by exactly the environment it must grow in.

Document Selection Decisions

Write down what you selected for and why, every generation. Without records, you cannot determine whether your selection is working, cannot explain the variety to others, and cannot adjust strategy when results diverge from expectations.

Selection Principles Summary

Mass selection — choosing the top performing individuals and mixing their seed — is the simplest and most widely applicable method, working best for high-heritability traits in cross-pollinating crops. Family selection reduces environmental noise by evaluating offspring groups rather than individuals, making it more powerful for complex traits. Multiple-trait selection can use index scoring, tandem rotation, or threshold culling. Selection intensity must balance improvement speed against diversity retention: moderate selection (top 20–30%) is appropriate for most survival contexts. Local adaptation is the ultimate goal — varieties shaped by your specific conditions over many generations will outperform commercial alternatives in your microclimate.