Talk:Heterosis

Talk

Discussion for: Heterosis • View • Links • Related changes

Use this page for verification notes, disputes, and action items. Keep narrative content on the article.

Page scope

Heterosis covers heterosis as an effect that can be measured: its reference points, its fall after the first generation, the explanations offered for it, its relation to divergence between parents, and what cannabis trials have measured. Loss of fitness on inbreeding belongs to Inbreeding depression, the making and use of lines to Inbred line, heterozygote advantage at a single locus to Overdominance, the relation between two alleles to Dominance and recessivity, the production of hybrid seed to F1 hybrid, crossing between taxa to Hybrid (biology) and the uniformity of selfed and hybrid lines to Outcrossing.

Data verification

Record source conflicts, figures that differ between studies and sources not used here. Keep narrative content on the article.

Discussion

Sources

Felsenstein (2019), Theoretical Evolutionary Genetics, pp. 85 and 411–414: the term, the rules for means of crosses and backcrosses, and the dominance and overdominance views.
Labroo and colleagues (2021), Frontiers in Genetics: mid-parent, panmictic-midparent, inbred-midparent, baseline and F₂ heterosis; the sign of heterosis across traits; proposed mechanisms.
Birchler and colleagues (2010), The Plant Cell: the case against complementation alone, with East (1936) and Duvick (1999) as they report them. Schnable and Springer (2013), Annual Review of Plant Biology; Chen (2013), Nature Reviews Genetics.
Flint-Garcia and colleagues (2009), PLoS ONE: heterosis across traits and genetic distance in maize. Fujimoto and colleagues (2018), Breeding Science: the history of the word and of hybrid corn and rice.
Krieger and colleagues (2010), Nature Genetics; Yu and colleagues (1997) and Hua and colleagues (2003), PNAS; Huang and colleagues (2016), Nature; Yang and colleagues (2017), PLOS Genetics: mapping and modelling in tomato, rice and maize.
Whitlock and colleagues (2000), Heredity, and Lynch (1991), Evolution: crosses between populations.
Mishchenko and colleagues (2022), Taurian Scientific Herald 125, in Ukrainian: seed-type and fibre-type crosses of selfed lines. Mishchenko (2017), Faktori eksperimental'noi evolucii organizmiv 21, in Ukrainian: combining ability. Mishchenko and colleagues (2025), the same journal, 37, in Ukrainian: male sterility and production.
Serkov and Uschapovsky (2024), Technical Crops 4(4), in Russian: Russian hemp hybrids. Dimitriev and colleagues (2022), IOP Conference Series: Earth and Environmental Science 981: heterosis in crosses of dioecious and monoecious hemp.
De Meijer (1994), Diversity in Cannabis, pp. 8–9: French and Hungarian programmes. Baluch (2026), in Achieving sustainable cultivation of hemp. Carlson and colleagues (2021), Journal of Experimental Botany. Barcaccia and colleagues (2020), Frontiers in Plant Science.
García-de Heer and colleagues (2026), Horticulture Research; McDonald and Lubell-Brand (2024), HortScience; Srikul and colleagues (2026), Agriculture and Natural Resources: trials of crosses in cannabis.

Where the sources disagree

Heterosis and inbreeding depression. Felsenstein calls a hybrid of inbred lines "in effect an observation of inbreeding depression in reverse". Labroo and colleagues limit strict opposition to baseline heterosis, and Whitlock and colleagues give heterosis between populations a different genetic basis from inbreeding depression within them. The two statements concern different crosses.
Heterosis in hemp production. De Meijer (1994) describes Hungarian F₁ hybrid cultivars in use. Serkov and Uschapovsky (2024) count five F₁ hybrids on the Russian register. Baluch (2026) describes hybrid hemp as more prevalent in the United States since 2014. Mishchenko and colleagues (2025) write that heterotic hybrids are absent from production because hemp has no known cytoplasmic male sterility. The last statement does not reconcile with the other three as a general claim.
Heterosis in drug-type breeding. Barcaccia and colleagues (2020) first call heterosis a driving factor for both fibre and drug-type breeding and later call it largely unexplored or undocumented in drug cultivars. The second statement concerns published evidence.
The reference for the seed figures. The abstract of García-de Heer and colleagues gives the rise of 3.9 to 155 per cent in seed yield against the midparents. Their discussion gives the same range against the midancestor values, and their Figure 5D shows the two references apart. The sentence on the article follows the discussion.
How much the F₁ exceeds the mother. Serkov and Uschapovsky give 1.5 to 2.2 times in their abstract and in their review of earlier work on over 100 combinations, and 2.1 to 2.5 times in their conclusions, for their own hybrids with a low-yielding mother. Their conclusions also ascribe the high seed yield to vigorous growth together with the predominance of females. Their review of earlier work says that the predominance of females counts for more.
What counts as heterosis. Labroo and colleagues define mid-parent heterosis as heterosis. Dimitriev and colleagues hold that the mid-parent overstates its frequency and size, and speak of heterosis only above the better parent. Mishchenko and colleagues report both, as hypothetical and true heterosis.
Dates for Shull. Fujimoto and colleagues date his finding to 1909. Felsenstein and Yu and colleagues cite his paper of 1908.
Who proposed dominance. Felsenstein names Jones (1917). Yu and colleagues date both hypotheses to 1908, dominance to Davenport and overdominance to East and Shull. Jones's paper concerns dominance of linked factors.
Epistasis in rice. Yu and colleagues (1997) and Hua and colleagues (2003) work from the same two parents, Zhenshan 97 and Minghui 63. The first found additive-by-additive interaction the commonest form and credited epistasis with a major role, on F₃ family means that halve dominance. The second measured heterosis in hybrids and held single-locus effects and dominance-by-dominance interactions sufficient for the F₁, and its authors allow that some additive-by-additive effects may be residual trait performance. The 86 per cent is for grains per panicle in 1998, assumes the effects are independent, and is not given for yield.
The correlation for total kernel weight. Flint-Garcia and colleagues print 0.56 in their Table 1 and 0.587 on their Figure 4A. The figures quoted in the article are from the table.
Mishchenko's year. Crossref's entry for the 2017 paper gives 1970. The journal's own header gives 2017, volume 21.

How far the evidence goes

Heterosis in cannabis has been reported from few crosses, and most comparisons differ in kind. Only Mishchenko and colleagues (2022) give mid-parent, better-parent and competitive heterosis together for one cross. García-de Heer and colleagues give midparent and midancestor values. McDonald and Lubell-Brand compare hybrids with selfed lines of each parent in two studies, with different pot sizes and dates, and give no mid-parent figure. Srikul and colleagues compare hybrids with the first selfed generation of each parent, and the parent plants themselves are not in the trial.
The parents in Mishchenko and colleagues' table are the varieties Afina and Ionino from which the selfed lines came, as the text says, and the hybrid is a cross of lines selfed four and six generations. The 9.29 g mid-parent in the table on the article is the mean of their 9.60 and 8.98 g; the percentages are theirs, and all three reproduce from their figures.
The dominance coefficient h_p in Mishchenko and colleagues describes where the hybrid's mean falls against the parents' means. Their value of 8.6 for seed mass is (11.96 − 9.29)/(9.60 − 9.29). A value above 1 is what they call inheritance by overdominance. It says nothing direct about overdominance at any locus.
Seed yield in crosses of dioecious and monoecious hemp mixes heterosis with the sex structure of the F₁. Serkov and Uschapovsky state this; none of these papers separates the two in figures.
Dimitriev and colleagues call their plants "cannabis-free" or "hashish-free" hemp and their hybrids "one sex". The title says monoecious hybrids, while the text describes dioecious mothers crossed with monoecious fathers, so the F₁ is almost entirely female. Their 38.5 per cent is 15.1 against 10.9 c/ha, and the higher parent, YSO-1, is also their monoecious standard. The parents are varieties, not inbred lines. Their Table 3 prints 143.9 against the mean of the parents, where 15.1 over the mean of 10.0 and 10.9 gives 144.5.
The per-plant seed figures of Dimitriev and colleagues do not say whether they compare plants of one sex. The authors note the view that an all-female F₁ enhances heterosis for seed and do not measure it.
The Thai trial is of two landrace accessions. Whether crosses between cannabis landraces in general show heterosis is not settled by it, and the height difference was measured under artificial light in a plant factory. The authors report no sign of inbreeding depression in the selfed generation, while its plants were shorter.
García-de Heer and colleagues grew 15 plants per line in 2-litre pots at high density, selected the most vigorous seedlings during inbreeding and suggest that the conditions may have limited heterosis. Their figures are not field yields.
The statement that Hungary was the only country where heterosis breeding of hemp became important is de Meijer's, of 1994, and the make-up of Kompolti Hybrid TC rests on a personal communication from I. Bócsa. De Meijer's account holds no figure for heterosis in that hybrid.
Mapping results in rice, tomato and maize are not cannabis. No locus for heterosis has been reported in cannabis in these papers.
Duvick's inbred-line result and East's conclusions appear on the article as Birchler and colleagues report them.

Terms

Mid-parent heterosis: the hybrid's excess over the mean of its parents; hypothetical heterosis in the Ukrainian papers.
Better-parent heterosis: the hybrid's excess over the better parent; true heterosis in the Ukrainian papers.
Competitive heterosis: the hybrid's excess over a standard variety.
Midparent and midancestor: in García-de Heer and colleagues, the means of the direct inbred parents and of the original open-pollinated lines.
General and specific combining ability: a parent's mean heterosis across its crosses and its value in one cross.
I₄, I₆ and S₆: a line after that number of selfed generations.

Images reserved in the article

Four images have places in the article but are not yet uploaded:

File:Cannabis-heterosis-harvest-traits-Garcia-de-Heer-2026.jpg (lead): Figure 5 of García-de Heer and colleagues (2026), Horticulture Research 13: uhag038, harvest traits of original lines, inbred lines and F₁ hybrids with midparent and midancestor values, CC BY 4.0.
File:Heterosis-four-measures-Labroo-2021.jpg (Measuring heterosis): Figure 1 of Labroo and colleagues (2021), Frontiers in Genetics 12: 643761, the diagrams and bars above its equations, CC BY 4.0.
File:Heterosis-excess-over-the-midparent-by-generation.jpg (After the first generation), a chart to be drawn from Felsenstein's rules: the excess over the midparent as 1 in the F₁, one half in the F₂ and F₃ under random mating, and one half, one quarter and one eighth in the F₂, F₃ and F₄ under selfing.
File:Maize-heterosis-and-genetic-distance-Flint-Garcia-2009.jpg (Divergence between parents): panels A to C of Figure 4 of Flint-Garcia and colleagues (2009), PLoS ONE 4: e7433, genetic distance and better-parent heterosis for total kernel weight, CC0.