Teosinte before the Holocene

Archaeologists and agronomists have accomplished a lot to understand the relationships of today’s domesticated crops and their wild progenitor species. But the people who domesticated plant species in the early Holocene were not working with today’s atmospheric CO2 and temperature regime.

That’s important when considering the effectiveness of early domestication efforts. The end of the last glaciation saw a worldwide increase in atmospheric CO2 level, and a general warming of the climate in the regions where agriculture first appeared. The wild progenitors of today’s crop species may have had very different productivity with the lower atmospheric CO2 and lower temperatures that characterized the climate 15,000 to 10,000 years ago.

Dolores Piperno and colleagues (2014) investigated the productivity of teosinte raised artificially in greenhouses with controlled CO2 and temperature regimes. They replicated the lower level of CO2 and lower temperature

It, therefore, becomes of considerable interest to ask if, during the late-glacial and early Holocene periods (c. 16–10 ka) when people first encountered, exploited, and cultivated many of the wild progenitors, the plants differed from modern wild populations, influencing crop plant evolution in ways that have been little considered. The last hunters and gatherers and first farmers worked with the phenotypes they saw, and it can be imagined they were attuned to and interested in the phenotypic variability they encountered on natural and cultivated landscapes. Unfortunately, chronologically-coarse and often geographically-uneven archaeo-botanical records do not adequately capture the range of phenotypic attributes that early cultivators experimented with. Moreover, the macrofossils (seeds, fruits, stems) that can best inform this question are often poorly preserved and have as yet to be recovered from Late Pleistocene and early Holocene records for many wild progenitors and earliest cultivars, including Zea (e.g., Piperno et al., 2009 and Piperno, 2011). Thus, modern representatives of crop plant ancestors constitute the basis for much of the morphological and genetic study of proto-domesticates and their wild ancestors.

Piperno and colleagues found two very interesting results:

  1. The teosinte was less productive under the low CO2, low-temperature regime that approximated the Late Glacial climate of Mexico.

  2. The teosinte raised under Late Glacial conditions had some phenotypic characteristics that resembled maize, that are typically not found in teosinte grown today or in those grown under Early Holocene conditions.

The second is the more newsworthy result. Teosinte is phenotypically very different from maize. The transformation of this wild crop into maize has been one of the most fascinating problems in studying the domestication of plants anywhere in the world. The idea that wild teosinte may have been more like maize in the past is pretty provocative.

This result is what had me reviewing the paper carefully. What the authors describe is a response rooted in phenotypic plasticity, in which a small proportion of the teosinte plants grown under Late Glacial conditions are more maize-like in some characteristics:

In every grow-out some of the plants in the chamber with late-glacial conditions (hereafter, referred to as the LGC) were complete maize-like phenotypes in branching architecture and inflorescence sexuality; like maize, they had a few, very short (non-measure-able) lateral branches tipped by female ears instead of tassels, that were attached directly to a single main stem tipped by a tassel (Fig. 3A) (hereafter, these plants with maize-type branching and inflorescence sexuality traits are referred to as “maize-like phenotypes” or MLPs). A total of six plants out of 33 from all the grow-outs combined representing every population studied had these characteristics (one plant from population 3 in 2009; one from pop. 4 in 2010; and one each from pops. 1 and 2, plus two from pop. 3 in 2011).
The maize-like phenotypes also exhibited a seed maturation strategy characteristic of maize, with most of their seeds maturing at the same time. In contrast, in other plants in the LGC and MCC, as in the wild, seeds matured sequentially from the tips of the branches to the base over a period of about two months, requiring several “harvest” trips to collect them before they began to fall off the plant soon after maturation.

Many models for the domestication of plants in the Old World have focused on the incidental effects of collecting. For example, many wild grain species have shattering ears – their grains easily pop off the ears to seed naturally near the mother plant. Those seeds would be much less likely to be collected by humans who were collecting the wild grains. If those people had extra seeds at the end of the winter and planted them, those seeds would preferentially represent plants with non-shattering ears. In this way, selection for non-shattering varieties could have been an incidental effect of human collection.

If Late Glacial collectors in Mexico could have selected incidentally on traits that were already present in the wild population of teosinte, the genetic variation that correlated with such phenotypic plasticity would likewise be subject to incidental selection from human collection.

Piperno and colleagues then conducted an experimental treatment in which the atmospheric conditions mimicked the Early Holocene, around 10,000 years ago. In this atmospheric treatment, the seeds of the maize-like phenotype plants from the earlier Late Glacial treatment continued to be more likely to produce maize-like phenotypes.

There are some unresolved genetic issues here. Today’s wild teosinte population likely does not preserve all the genetic variation that once existed across the range where maize was domesticated. The original local variants that gave rise to the first maize may therefore have had phenotypes lacking in today’s wild teosinte. When we look at the variability of today’s teosinte in the face of atmospheric variation, then, we might be looking at an axis of phenotypic variation that falls short of the phenotypes of teosinte during the Late Glacial.

The other major finding of the study is that the plants were less productive under the Late Glacial atmospheric treatment. Also found by other studies, that has become a key observation about the productivity of wild plants in the time leading up to domestication.

People became intensive collectors in some parts of the world in the wake of the Last Glacial Maximum. Yet the domestication of crop plants and transition from collecting to agriculture did not happen until substantially later. Why the delay? Social factors don’t seem a sufficient explanation: Some areas had a long history of sedentism and relatively high population density before domestication, others did not.

One answer is that the productivity of some wild grains increased at the transition to the Holocene, making it more possible to reseed from collected stores, and less necessary to depend on a broad array of less-preferred plants. So agriculture may have had not only a general environmental trigger – the end of the last Ice Age – but also a more specific environmental trigger: the phenotypic plasticity of wild plant species in the face of atmospheric conditions.


Piperno DR, Holst I, Winter K, McMillan O. 2014. Teosinte before domestication: Experimental study of growth and phenotypic variability in Late Pleistocene and early Holocene environments. Quaternary International (in press) doi:10.1016/j.quaint.2013.12.049