A Wikiblog E-Book by Norman Uphoff with many others
Chapter 6: NUTRIENT UPTAKE AND PATTERNING OF CROP GROWTH
Once the significance of having prolific root systems and more abundant life in the soil was understood, as discussed in the two preceding chapters, it was clearer why the practices that Fr. Laulanie had assembled into the System of Rice Intensification could raise the productivity of the resources, the land, labor, water, seed and capital that farmers invested in growing rice.
Wider spacing between rice plants provided more room for their roots and their canopies to grow. Keeping the paddy soil moist but mostly aerobic encouraged the growth of both plant roots and beneficial organisms in the soil. Supplying the soil with more organic matter provided energy for the myriad miniscule and many visible creatures that enhance soil systems’ functioning, and that made the soil better structured for the rice roots to grow in and spread. All of these effects were enhanced by the active soil aeration that resulted from SRI’s mechanical weeding.
This made good sense. But how did the transplanting of very young seedlings come into the picture? From the factorial trials discussed at the start of the next chapter, we learned that this was the SRI practice the contributed the most to higher yields. And to explain the profuse plant growth, one had to account for where the soil nutrients that supported such growth come from? Skeptics insisted that the soils around Ranomafana did not contain sufficient nutrients to support such growth, at least not without applying synthetic fertilizers.
From understanding the growth of rice plants in terms of phyllochrons, mentioned in Chapter 3, we were able to gain considerable insight into how and why SRI methods succeed as they do, just as Fr. Laulanié had come to understand this once he read Didier Moreau’s book on rice improvement in 1987. Knowing more detail about the sequencing and timing for tillers and roots to emerge from the seed and ensuing seedling did help to explain why farmers should transplant rice seedlings when these are still very young, less than 15 days old rather than being 3, 4, 5 weeks old, or more. This is elaborated on in the last half of this chapter.
But just planting young seedlings will not confer much advantage as an isolated practice. Plants need to access and take up nutrients from the soil in order to grow and prosper. Thus, understanding the connections between plants’ uptake of nutrients and their associated patterns of growth was essential for grasping the effects of SRI crop management.
In Chapter 4, it was noted that rice plants’ uptake of nitrogen, and by extension of other nutrients, is a function more of their demand for N than of the supply of N that is in the soil around the roots. This counterintuitive relationship was reported but not much considered in the agronomic literature. Rice plants appear to have evolved innate regulatory mechanisms whereby their roots will not take up nutrients in excess of the plant’s requirements.
Indeed, plant roots, it was known, will exude nutrients rather than absorb more than the plant needs. This called into question the Green Revolution strategy of trying to increase crops’ growth by plying them with ever-greater amounts of chemical fertilizer, having planted crop varieties that had been bred to be able to utilize inorganic nutrients more efficiently.
I read the scientific articles discussed below at about the same time that I was learning about the research results from Mexico reported in Chapter 4 which showed that doubling the supply of nutrients provided to young maize plants, beyond what plant nutritionists calculated they needed, did not result in any enlargement of the plants’ roots or above-ground canopy.
It was understandable that plants’ roots might not grow larger when provided with an excess of nutrients because there was little need for roots to grow if plenty of nutrients are at hand. What was not expected was that with a super-abundance of nutrients available, their supply having been doubled, the plants’ shoots above ground did not increase in size; instead they were reduced by one-quarter. An excessive supply of nutrients stunted the plant’s growth.
When looking for further research on this relationship, I found that almost all plant and soil research dealt with the negative effects of nutrient deficiencies in the soil, including stunting and greater susceptibility to pests and disease. It was hard to find studies that delved into the effects of nutrient excess, beyond observing that an excessive supply of nutrients could cause unproductive vegetative growth and also disease problems, attracting pests, and possibly impeding plant growth because of toxicity. Researchers seemed not to care about the effects of having too much nutrient availability, or why plants did not benefit from a supply of nutrients exceeding their requirements.
What was most thought-provoking in the research from Mexico reported in Chapter 4 was that when the growing plants were provided with a greatly reduced but continuously-available supply of nutrients (50 times less), their shoot growth equaled what they achieved with the full recommended amount of nutrients. This normal growth was accompanied by an eight-fold increase in the plants’ root volume! This growth response of the roots enabled the plants to extract sufficient nutrients from the much-diluted solution to achieve what would be considered ordinary growth.
These numbers challenged the standard view of a roughly proportional relationship between nutrient provision and plant growth. The relationship was anything but linear. Excessive nutrient supply was seen as something wasteful, but not necessarily deleterious. These findings reported by Ana Primavesi in her magisterial book on tropical soils attracted little attention, however, perhaps because the numbers derived from formal experiments were at odds with accepted agronomic thinking and did not come from a prestigious institution.
The specific numbers from similar experiments will surely vary for different crops and different circumstances. But this research clearly indicated that plants do not grow according to preordained genetic instructions. Plants have agency and act on behalf of their growth and reproduction objectives, responding to and adapting to variations in their environment. The growth process itself is exceedingly complex, more contingent and compensatory than predetermined. Plants’ size and grain production is not a direct, proportional result of the amount of external nutrients that are supplied through fertilizer. Plants are not animate carbon-based ‘machines,’ to be designed and redesigned for our purposes.
The articles sent to me in 1999 by Guy Kirk after my presentation on SRI to IRRI rice scientists at Los Baños in the Philippines (Chapter 4) helped to make sense of what we had seen in Ranomafana. In 2002, after reading these and other scientific articles, I put together a paper on the relevant relationships among plant growth, roots, nutrients and phyllochrons that I could find reported on in the literature. This paper was circulated among SRI colleagues but was not submitted for publication at the time because we did not have enough experimental evidence to satisfy journal reviewers. The following section summarizes what was concluded from the ‘mainstream’ scientific research available two decades ago.
THE DYNAMICS OF HOW RICE PLANTS TAKE UP NITROGEN FROM THE SOIL
It has been widely believed and often stated that nitrogen is the most important factor affecting rice growth and yields. Kronzucker et al. wrote in 1998 that in the irrigated lowland rice areas of Asia, "N is generally the main factor limiting the realization of yield potentials." Concurrently, Ladha et al. asserted: "To increase grain yields, additional N must be applied as fertilizer." Yet these latter authors also added, surprisingly: "The use of fertilizer-N has increased with time, but the yields have often remained constant in both experimental and farmers' fields."
This observation contradicted what the authors had just written and the widely-held opinion that applying more N fertilizer to fields is the best and most certain way to increase crop yield. Was the injunction to increase fertilizer use more a matter of belief than scientific fact, though? Was it like the assertion that rice plants that have more tillers will have fewer grains per tiller? Or like the mistaken idea that rice plants grown in standing water will be more productive? What we were learning from SRI experience and from scientific research was confirming that although rice plants can tolerate continuous flooding, this inhibits their highest yield.
It was evident that there are diminishing returns to applying more nitrogen fertilizer: additional inputs of N fertilizer were giving less and less increase in output. In China, at the start of its Green Revolution in the mid-1960s, applying an additional kilogram of N fertilizer could add 15-20 kg to yield. By the start of the 21st century, this ratio had fallen to where only about 5 kg of additional rice being produced per hectare in response to the application of one more kilogram of fertilizer. This was a four-fold decline in fertilizer’s productivity.
Worldwide, nitrogen-use efficiency (NUE) for cereal production (wheat, maize, rice, barley, sorghum, millet, oats and rye) was said to be approximately 33%. This means that about two-thirds of the fertilizer that farmers put onto their fields was not being taken up by crop plants. With irrigated rice, N-use efficiency was estimated to be as low as 20-30%, the rest being lost into the atmosphere through volatilization or into ground water supplies through leaching.
It was a fair question to ask whether these diminishing returns and such low efficiency of nitrogen use might be a consequence of the way that rice plants were being managed, along with the soil, water and nutrients that they need, rather than being due to some intrinsic qualities of the plants or the soil systems?
Rice plants certainly need to have ample availability of nitrogen and other nutrients, and they benefit from taking up more N if they have enough of all the complementary nutrients that plants need to carry on metabolism and to synthesize carbohydrates, amino acids, and other organic compounds. However, the research cited above suggested that plants’ uptake of some nutrients was, at least in part, a consequence of their growth rather than just as the cause for such growth.
This idea was consistent with the low use-efficiency of externally-provided nitrogen that was widely reported, and with the diminishing returns seen when N fertilizer was supplied in large amounts. This had to be discomforting for proponents of Green Revolution strategies.
Given the low use-efficiency of inorganic N fertilizer, Ladha, Kirk and other scientists at IRRI and other institutions spent considerable time and effort during the 1990s studying biological nitrogen fixation (BNF), which is driven by microbial activity, as a way to get rice plants to take up more N so that they would produce more grain. This interest was prompted partly by the fact that "excessive uptake of fertilizer-N leads to increased risk of disease and to lodging."
The observation that heavy applications of inorganic nitrogen caused problems for growing plants, like greater vulnerability to pests and disease and falling down (lodging), supported the proposition that rice plants need and benefit from balanced diets of nutrients, micronutrients as well as macronutrients, rather than just from having more N or more phosphorus (P) and/or more potassium (K).
We were finding in the field that SRI-grown rice crops, with little or no inorganic fertilizer, seldom experienced lodging and had fewer losses from pests and disease. Moreover, the high yields achieved under SRI management led us to think that biological N fixation could become a significant source of N when the rice plants are grown in well-aerated soil with supportive plant management practices.
We took note of a report by Ladha et al. that N uptake and grain yield generally increase with greater N supply only up to some optimum point, where yields reach a maximum of 8 to 10 t/ha, and beyond this, yields decline with increased application of N fertilizer. This was an acknowledgement that excessive nutrient supply (as judged by the plants) is counter-productive. They wrote, "There was no significant increase in yield beyond 150 kg/ha [of fertilizer application],” and in fact, "Yields for multiple varieties peak out at about 8 t/ha, even with high N applications, up to 200 kg/ha."
These relationships that these rice scientists were reporting could reflect the conditions under which the crop was being grown rather than represent the full genetic potential of the rice planted. Data from Madagascar and elsewhere were showing that SRI yields of 8 tonnes per hectare to be more an average than a maximum. If continuous flooding caused rice plants’ root systems to degenerate, this would surely constrain crop performance.
We thus doubted that the reported ‘optimum’ point of N-fertilizer application, up to 150 kg of N per ha, would give ‘maximum’ grain yield. Why? Because we were so often seeing yields higher than 8 to 10 tonnes per hectare in Madagascar and elsewhere, and with the best use of SRI management practices, rice yields could be twice this much.
SRI results indicated that these plants relying mostly on organic sources of nutrition were taking up more N than they did when supplied with inorganic fertilizer. Or maybe SRI-grown plants did not need as much N as had been calculated based upon conventional management practices when rice plants are crowded and flooded. Or maybe rice seedlings transplanted at a very young age grow into plant phenotypes that are inherently more efficient and productive because of the conditions for plant growth that SRI created. Or some combination of these effects.
It was thus interesting to read the description of root-nitrogen relations that Kirk and Bouldin had offered some years before. Rice roots' uptake of N increases as the concentration of N in the root zone increases only up to a certain level or plateau, they wrote, and beyond this point, "further increases in concentration do not appreciably increase the uptake rate." Rice roots' uptake of N, they proposed, is strongly dependent: (a) on the rate of plant growth [its demand for N], (b) on the ion concentration [of N and other nutrients] within the root, and (c) on the conditions of the plant as a whole.
This latter consideration "impl[ies] a mutual regulation of root and shoot activities," their article said, indicating that a rice plant’s shoots will not emerge if its roots cannot grow, and vice versa. Both roots and shoots emanate from the same meristematic tissue at the base of the plant. Because several interdependent processes affect rice plants’ nitrogen uptake, Kirk and Bouldin concluded that plant roots’ intake of N is "independent of the concentration [of N] at the root surface per se." This was a thought-provoking conclusion.
Kirk and Bouldin went on to explain that when the rice plant does not have a current need for N, that is, when its internal N status is satisfactory, its roots down-regulate their intake (or uptake) of N. Indeed, when the plant has a sufficient supply of N, its roots will exude as many N ions as they take in, they wrote.
Conversely, when the plant has a need for more N, it up-regulates the functioning of its roots so that they take in more nitrogen. Almost everyone knows the maxim: “You can lead a horse to water, but you cannot make it drink.” This could be extrapolated to rice, suggesting that rice is like a horse that can be ‘led’ to nitrogen, but it cannot be made to ‘drink’ more nitrogen than it needs.
This insight meant that the Green Revolution strategy for raising crop yields by increasing the application of fertilizer was self-limiting. Providing nutrients for plants could raise crop production to some extent when there were nutrient deficits to be filled. But forced-feeding plants would be counter-productive because of evolved regulatory mechanisms that limit plants’ capacity to take up more nutrients than are needed for their growth.
The Green Revolution strategy for agricultural improvement is reminiscent of the farming practice in France where, to make ducks and geese gain weight more quickly, large amounts of grain are literally stuffed down their throats, with the aim of making their livers become larger and fattier, so that more and tastier paté de foie gras can be made for French consumers.
That only a fraction of the inorganic nitrogen being applied on rice-crop soils was being taken up by the plants was certainly economically costly and inefficient. But it also was (and is) environmentally-damaging because it raises the levels of nitrate in surface and groundwater supplies.
Similarly, the runoff of phosphorus from agricultural fields from overuse of phosphate fertilizers is deleterious for the natural environment because this pollutes water bodies. Excessive concentrations of P as well as N cause eutrophication and algal blooms in rivers, ponds, lakes and oceans. The widespread supply-driven strategy of seeking to expand crop production by increasing the provision of fertilizer thus should be re-examined and revised except where there is evidence of nutrient deficiencies because otherwise it has environmental as well as economic costs.
SRI PLANTS TAKE UP MORE NUTRIENTS AND CONVERT MORE INTO GRAIN
Having been recognized as the most outstanding college graduate in Madagascar in 1998, Joeli Barison, whose baccalaureate thesis research was described in Chapter 3, received a USAID scholarship to study at Cornell for a master’s degree in crop and soil science. His thesis research with both on-farm and on-station data and supervised by Erick Fernandes investigated further various mechanisms for the effectiveness of SRI methods. Data were gathered together with a fellow Master’s student in agricultural engineering, Oloro McHugh, making the empirical basis of their respective theses quite extensive.
Joeli and Oloro analyzed data obtained from 109 farmers who grew rice in four locations of Madagascar using both SRI and traditional methods of cultivation on their respective fields. This research design controlled for the effects of differences in soil, climate, and farmer skill. Joeli also conducted controlled trials to make precise crop measurements at a research station near Beforona.
The farmer-field comparisons showed that SRI methods with compost gave yields of 6.26 tonnes per ha, while the same farmers on the same soils with their usual methods averaged 2.63 tonnes per ha. With use of recommended ‘modern’ methods including fertilizer application, the farmers produced an intermediate yield of 4.92 tonnes per ha.
The on-station trials showed an even greater effect of SRI methods on rice roots’ growth than Joeli had found in his previous study in Ranomafana. Root-pulling resistance (RPR) for single SRI plants averaged 55.2 kg, while clumps of three conventionally-grown plants on average required 20.7 kg to pull up. This amounted to 6.9 kg per plant, which was eight times less than for an SRI plant. Rice plants grown with farmer or ‘modern’ methods, where fields were kept flooded, had most of their roots in the top 20 cm of soil, whereas the roots of SRI plants had 2-3 times more biomass and extended to depths below 30 cm.
Joeli’s data on rice plants’ nutrient uptake were most illuminating. Rice grown with SRI methods took up nitrogen into their above-ground biomass at a rate of 95.07 kg of N per ha. The same variety of rice on the same soil but grown with farmers’ methods accumulated 49.99 kg of N per ha. SRI phenotypes thus took up 91% more nitrogen per hectare. Joeli’s data showed a similar 66% advantage of SRI rice plants for the uptake of phosphorus, and an even greater advantage for potassium.
These differentials could have resulted from several factors. There could have been more nutrients available in SRI plants’ root zones from the provision of compost and associated aerobic biological activity. Or possibly SRI plant phenotypes were able to take up more N, P and K and/or were better able to utilize these nutrients in the above-ground organs. Such differences might account for much of the disparity between SRI and farmer-practice yields. However, the research was not designed to sort out these effects.
Joeli’s analysis of the data revealed something very interesting about the phenotypical differences between SRI and conventionally-grown rice plants. The differences seen in the graph shown below from Joeli’s thesis have not been fully explored and explained. He analyzed the data that he had gathered using what is called the QUEFTS model (Quantitative Evaluation of the Fertility of Tropical Soils) adapted to the study of nutrient dynamics in rice. This model assesses plants’ internal use-efficiency of nutrients, considering interactions among the three major nutrients, N, P and K. As the same varieties were being compared, the differences would reflect the effects of crop management and phenotype differentiation rather than genetic differences.
With this model, Joeli regressed the respective plant uptakes of N, P and K on the measured grain yield, assuming a parabolic, non-linear relationship. The respective regression lines for plants grown with either SRI methods (the upper line in the figure below, representing the distribution of blue diamonds) or conventional methods (the lower line, approximating the relationship among the purple triangles) showed that as more macronutrients were taken up by the plant, the increments to grain yield declined more rapidly in the conventional rice plants than in the SRI plants. The height and slope of this regression line indicated a more rapid decrease in conventional plants’ internal efficiency for utilizing N, P and K when converting these nutrients into grain, and greater efficiency on the part of SRI plants. The figure below shows this relationship for nitrogen. The corresponding graphs for phosphorus and potassium are not shown here because they were essentially the same as for nitrogen.
The two regression lines show a much steeper increase in grain yield as a function of the uptake of N for the SRI plants compared to the N uptake by rice plants grown conventionally. Similar analysis showed the same greater yield responsiveness of SRI plants to the uptake of P and K. Under crop management by the same farmers on the same soils, the yield of conventionally-grown rice plateaued between 4 and 5 tonnes per ha, while the yield from SRI plants was still on an up-curve within the range of plants evaluated.
Since these data were from crops grown on farmers’ fields, their fertilization was not uniform and controlled as it would have been with on-station trials. However, most of the farmers surveyed for this study were not applying either compost or fertilizer on either of their plots, SRI or conventional. What was being measured was thus not fertilizer application but what the plants were taking up from the soil. As yields were measured with the same methods for both SRI and conventionally-grown rice, possible errors in measurement would not explain the divergent trajectories for rice crop yield response under the respective practices.
Since no study was done of metabolic processes, we cannot know what was different about the physiology of the respective phenotypes, although from studies reported in Chapter 11 we know that SRI plants have specific advantages like higher levels of chlorophyll in their leaves and faster rates of photosynthesis to synthesize carbohydrates which are then accumulated in their grains.
An important hypothesis still to be tested is that given the differences in root system growth and functioning, with more beneficial microorganisms in the rhizosphere and in the plant itself, there will be more uptake of micronutrients in SRI plants. There is already some evidence of this reported in Chapter 11. With more uptake of micronutrients, SRI plants would be working with what would be called in the language of playing-card games ‘a full deck,’ one not lacking a number of important cards, i.e., micronutrients.
Plants that are abundantly supplied with N, P and K but not with compost or other organic soil amendments would have plenty of macronutrients relative to their supply of micronutrients. The latter are needed to synthesize enzymes and other compounds that are necessary for rapid and efficient growth. Soils well-supplied with organic matter will have more of the essential trace minerals like boron, zinc or copper which plants need for their metabolic processes and for their immune system to function. Plants that have a sufficient supply of the full range of the micronutrients that they need for growth and health could produce more grain per unit (molecule) of macronutrient taken up, as suggested in the figure above.
Having an ample supply of both macro- and micronutrients is of course desirable, but as discussed above, plants’ uptake of nutrients is not driven by supply so much as by demand. For plants’ uptake of nutrients and growth to proceed efficiently, plants need to have sufficient amounts of all the relevant nutrients. This brings us back to the need to understand the processes of plant growth. Here the story of SRI intersects with the concept of phyllochrons that, as mentioned in Chapter 3, helped Laulanié to understand why his transplanting young seedlings had such a great and demonstrable impact on rice plants’ growth.
The idea of phyllochrons provides an empirical basis for the proposition of inverted causation, that plants’ uptake of nutrients should be regarded more as a consequence of their demand for nutrients than as being mostly a matter of nutrient supply. This paradoxical relationship would not have occurred to us if we had not so often observed high yields of rice resulting from the use of SRI methods on soils that were measurably very low in their stock of nutrients. These yields surpassed what some scientists considered as a ‘biological yield ceiling’ for rice. This ceiling, it turns out, was more dependent on management than on nutrient supply.
PHYLLOCHRONS IN RICE: WHY TRANSPLANTING SEEDLINGS AT A YOUNG AGE ACCELERATES PLANT GROWTH
The careful transplanting of very young seedlings is important for SRI success because rice plants’ production of tillers with associated leaves and concomitant downward growth of roots is a well-structured, cumulative process. It is not simply a quantitative process as implied by agronomists’ reference to there being periods of low, moderate and high tillering during the plants’ vegetative growth phase. A first phase of vegetative plant growth, during which tillers and roots emerge, precedes a second phase of reproductive activity, when many if not all of the rice plant’s tillers flower (this period of flowering is known as anthesis), and their panicles begin the successive processes of grain formation, grain filling, and grain maturation.
Tillering proceeds according to a genetically-determined pattern that can be analyzed in terms of phyllochrons, a patterning common to all plant species in the grass family (gramineae). How rapidly or how slowly this process of tiller and root emergence proceeds, and with what consequences for the number and size of plant organs, is quite variable. It depends greatly upon the conditions under which the plant is growing, enumerated below. Transplanting seedlings at a time when there will be the least trauma to the plants, and especially to their frail roots, will have a beneficial effect on their subsequent growth and productivity.
This patterning was first recognized and documented by a Japanese scientist, T. Katayama, who during the 1920s and 1930s meticulously studied these patterns of tillering and root growth in rice, wheat and barley. His first technical paper on this subject was published in 1931, in Japanese with a brief summary in English, and his findings were finally presented more fully in a book published 20 years later, after the Second World War was over. Because this book was never translated into English, its substance is not much-known outside of Japan, although a few discussions of phyllochrons have become available in English.
Rice scientists in Western countries have had little acquaintance with the concept of phyllochrons, being satisfied with the cruder notion of ‘growing degree-days,’ which is concerned only with weather. The emergence of SRI on the world scene over the past 20 years has helped to bring the concept forward in rice science circles. I remember well how shortly after I learned about phyllochrons from Laulanié’s writings, I consulted the Oxford University Press’ Dictionary of Plant Sciences (1998 edition). There was no entry on ‘phyllochrons’ to be found within its 600+ pages.
When checking on the internet in 2000, I found that most of the published research on phyllochrons applied the concept only to wheat or to forage grasses (in Australia). Fortuitously, there was whole issue of the journal Crop Science devoted to papers on phyllochrons which I found. These papers had been presented at a symposium held in November 1993 at the annual meetings of the Crop Science Society of America. Most of the papers in that special issue reported on research on phyllochrons in wheat, however. Only one paper was on phyllochrons in rice, contributed by Japanese researchers.
In his 1993 article, Laulanié discussed the significance of understanding phyllochrons for increasing rice production. He noted that transplanting trauma to rice seedlings could be minimized by relocating them from their nursery to the main field before the start of their fourth phyllochron of growth. This is usually about 15 days after seedling emergence, depending on temperature and other conditions, and when there was less root growth to disturb.
If the plant experiences trauma after its primary tillering and associated root proliferation begins, the plant’s growth trajectory will be slowed, and the plant will be unlikely to complete more than 7 or 8 phyllochrons of growth before its reproductive phase begins, superseding the phase of vegetative growth.
If all of the conditions for plant growth are optimal and the young seedling has not been traumatized during transplanting, it should be able to complete 10 or 12 phyllochrons of growth, or maybe even more cycles of tiller emergence and root growth, before it begins to form grains and fill them with carbohydrates. Each phyllochron interval will have an increasing number of tillers and roots emerging, as shown below.
The stump of an SRI rice plant shown below was given to my wife and me by SRI farmers in East Java, Indonesia in 2009. It has 223 tillers, all grown from a single seed. Usually individual rice plants have between 10 and 30 tillers. This large a number of tillers means that the plant was able to keep producing stalks and panicles well beyond its 12th phyllochron of growth, that is, before it ended its phase of vegetative growth and entered its reproductive period of flowering, after which it produce seeds and nurtured them to maturity.
Phyllochrons, as noted above, are variable periods or intervals of plant growth that are observable in all of the gramineae (grass) species. Each phyllochron is a period during which tillers and roots are formed, with these organs (structures) emerging from the same meristematic tissue at the base of the plant.
Tillers grow upward and roots downward as a functional unit called a phytomere, which includes also a leaf that accompanies the tiller. In rice plants, the duration of phyllochrons is usually around 5 to 7 days, but often longer or occasionally shorter, depending how adverse or favorable the growing conditions are.
During each successive phyllochron, larger and larger numbers of phytomers (tillers and roots) emerge, as shown in the diagram below. This diagram constructed by Laulanié on that assumption that 12 phyllochrons of growth would be completed during the vegetative growth phase before the reproductive phase began.
If the plant’s growing conditions are good, that is, favorable for rapid cell division, elongation and differentiation of cells, its ‘biological clock’ runs more quickly and smoothly, to use the metaphor of horology. With good conditions, each phyllochron will last about 5 days as measured by calendar time, although what is important is not the passage of time so much as the completion of certain biological processes.
If phyllochrons are relatively short, the plant can complete more cycles in which one or more phytomers, eventually many, emerge from the plant’s base, before its period of vegetative growth ends. Then tillers and roots stop emerging, and some, many or most of the plant’s tillers will flower and form panicles, beginning the plant’s reproductive period.
Conversely, if a plant’s growing conditions are not so favorable, its ‘biological clock’ will run more slowly. Its phyllochrons will be longer, and fewer cycles of growth will be completed before the plant’s vegetative phase ceases, and flowering sets in. The length of phyllochrons will vary according to climatic, soil, varietal and other influences as discussed below. Under ideal conditions for growth, phyllochrons can be as short as 4 days, whereas with unfavorable conditions, they can be as long as 10 days or more.
If a rice plant completes 11 or 12 phyllochrons of growth before it commences reproduction, dozens of phytomers (units of tiller, root and leaf) will be produced in the final cycles of growth. A rice plant that completes only 7 or 8 phyllochrons of growth before flowering, on the other hand, will have only 8 to 13 tillers. A plant that can complete 12 phyllochrons before beginning panicle initiation can have as many as 84 tillers, with a corresponding increase in its number of roots.
A schematic representation of tillering patterns organized in terms of phyllochrons, as worked out by Fr. Laulanié from Katayama’s analysis, is shown below. Rice plant roots, which grow below-ground in a similar pattern in time and space, are not shown.
The main tiller and main root, indicated by the vertical line in the center of the diagram, emerge from the seed during the plant’s 1st phyllochron of growth, usually sprouting between 5 and 8 days. Then there is no further emergence of tillers and roots during the 2nd and 3rd phyllochrons. This explains why it is the advantageous to transplant rice seedlings early, during this period when there will be least disturbance of their growth trajectory.
During the 4th phyllochron, another tiller and root will emerge from the base of the main (first) tiller and root, and then also another tiller and root in the 5th phyllochron. (These are called the 1st and the 2nd primary tillers and roots). Thereafter, the tillering and associated root growth become more complicated and accelerated, proceeding according to what is known in biology and mathematics as a 'Fibonacci series' -- 1, 1, 2, 3, 5, 8, 13, 21, 34, etc.
In such a numerical progression, the new growth in each period after the first period is equal to the sum of the growth in the two preceding periods. In rice plants, beyond the 10th phyllochron of growth, there appear to be physical space constraints on tiller emergence that make the Fibonacci series from this point on more approximate than exact, but the principle of growth holds.
The diagram below is Fr. Laulanie’s representation of rice plants’ tillering beyond the 4th phyllochron, simplifying the scheme that had been worked out by Katayama, translating Katayama’s extremely complicated table into a more comprehensible drawing. The graphic makes more sense when coupled with the table that the priest worked out from Katayama’s analysis and from his own observations, also shown below.
With conventional rice-growing practices, close spacing of plants and hypoxic soil conditions plus transplanting more mature seedlings (after the start of their 4th phyllochron of growth) combine to slow down the biological clock of rice plants, so that they do not achieve their maximum potential for tillering.
SRI plants with over 100 tillers, and there are some, have extended their phase of vegetative growth beyond their 12th phyllochron, thanks to ideal growing conditions. The Indonesian SRI rice plant shown above would have had a prolonged phase of vegetative growth reaching into its 15th phyllochron. Unfavorable growing conditions, conversely, will slow down the ‘clock’ so that only 7, 8 or 9 cycles of tiller and root emergence occur before the plant flowers. At this point, the plant’s resources are switched from vegetative growth into grain formation and grain filling.
WHAT REGULATES RICE PLANTS’ BIOLOGICAL CLOCK?
When rice seedlings are transplanted after the beginning of their 4th phyllochron, and certainly when seedlings are transplanted several phyllochrons later than this, the mature plants will not achieve their full growth and full yield potential. This physiological limitation will be compounded if the seedlings have been traumatized during transplanting, and if hypoxic soil conditions lead to root degeneration. Root growth, tillering, and grain filling will all be reduced by any sub-optimal growing conditions. More favorable conditions, on the other hand, enable rice plants to complete a larger number of phyllochrons of growth (cycles) before their panicle initiation. This is important because as seen above, both tillering and root growth are accelerating processes.
The concept underlying this analysis is that rice plants' growth will proceed according to some kind of 'biological clock' that runs faster or more slowly depending on the total effect of favorable and/or unfavorable growth conditions. The ‘clock’ is regulated operationally by the speed with which the plant’s cells are growing, elongating and dividing, resulting in cells’ more rapid growth, elongation, and division which cumulate into the growth of organs and plants.
As seen above, the hypoxic soil conditions under which most irrigated rice is grown, as well as the usual management practices employed, will reduce the rate of rice plant growth. This in turn slows the plant’s demand for and uptake of N and other nutrients. Such an understanding of rice physiology makes clearer why yields do not rise much, beyond a certain point, by increasing the application of N fertilizer.
A de facto ceiling on rice yield is created by current practices including crowding and flooding. The reported yield limit of 8-10 tonnes per ha cannot be breached by providing plants with more N than they need and will absorb. If rice plants are given more N than they need, their roots will down-regulate their rate of nutrient uptake, and the plant’s growth is inhibited. This kind of mechanism appears to be operative in the results of the experiments with maize seedlings reported in Chapter 4.
What speeds up or slows down the rice plants’ biological clock? Understanding this can help farmers create the best conditions for their crops’ growth. While this has been worked out for rice production, the same agronomic factors should help improve the production of other crops as well. Indeed, this has been seen with many other crops (Chapter 14).
In their paper in Crop Science on rice phyllochrons, Nemoto et al. reviewed the factors that Japanese rice scientists had found to be influencing plants’ biological clock, making it run faster or slower. This rate accordingly shortens or lengthens the plants’ phyllochrons, the cycles of tiller and root emergence through which rice plants pass before they re-direct their internal resources from vegetative growth to reproduction and grain production.
The seven positive factors, and their converse inhibiting factors that slow down the biological clock, can be summarized as follow:
Warmth is beneficial, although only up to a certain point because temperatures that are too high will slow and eventually stop growth. Conversely, cold temperatures inhibit plant growth and can even stop or prevent it.
Sunlight is needed for plant growth, obviously. But while this is correlated with warmth, it is a different factor because it relates to photosynthesis rather than to the plant’s metabolism. Just as sunlight is beneficial for plant growth, shading is detrimental.
Spacing is something to be optimized for best production from a given land area, although individual plant growth is maximized by wider spacing. Obviously when plants are crowded together and they must compete for light, water and nutrients, this will inhibit their root growth and their tillering.
Plants slow their growth (their biological clock) when they are close to each other, sensing somehow that density will be disadvantageous for themselves and their neighbors if they grow more than their situation allows in terms of available light, water and nutrients. This will be discussed more below.
Nutrient availability is a stimulus to plant growth provided that the plant can make productive use of the nutrients. With nutrient deficiency, on the other hand, plants are not only deprived of building-blocks for their growth, but they slow their growth so as not to become so large that their metabolism cannot be sustained, meaning that they would die without achieving their goal of reproduction. This will also be discussed below.
Soil structure should also be favorable, that is, loose, friable and crumbly, so that plant roots can penetrate it and the root system can expand easily, accessing a larger volume of soil from which the plant can acquire both nutrients and water. If the soil is compacted and not porous, root growth is impeded. Tillering and leaf growth are at the same time also inhibited because there is a high degree of symmetry and reciprocity between the plant’s organs above- and below-ground.
Soil aeration is important for both plants and for the aerobic microorganisms on which they depend. Both need to have sufficient oxygen. Soil porosity and aeration are beneficial also for venting undesirable gases in the soil like carbon dioxide (CO₂) and hydrogen sulfide (H₂S). These can build up and become toxic with no ventilation. Soil organisms, as discussed in Chapter 5, improve the porosity (air spaces) in the soil, so that both air and water can diffuse better throughout the soil.
Water is as important as air for plants and beneficial soil microorganisms. These last two factors are inversely related since for any given amount of pore space in the soil, having more water or air means having less of the other. All plants need some water, but when there is too much water in the soil, there will not be much oxygen. On the other hand, when the soil is desiccated, there will be plenty of air, but little or no water.
One of the insights gained from phyllochron theory and from SRI experience is that instead of thinking in terms of competition between plants, it is more useful to think about their cooperation. The observed behavior of rice plants (understood in terms of what speeds up or slows down their biological clocks, thereby shortening or increasing the length of phyllochrons) suggests that plants have not evolved according to algorithms that maximize their individual growth so much as according to algorithms that regulate and optimize this. The growth of rice plants’ roots and tillers is inhibited when they are crowded together.
The goal of rice plants is not to produce the largest possible harvest of grain. That is humans’ objective. In anthropomorphic terms, rice plants strive to survive through their entire cycle of growth and development so that they can reproduce themselves. That is how plants have survived on this earth for over 400 million years. Theirs is a symbiotic strategy that includes coexistence with microorganisms, most of which (but certainly not all) help plants to survive and while benefiting from plants’ existence. The plants themselves also function somewhat symbiotically. These relationships can be sketched schematically as shown below.
FEED AND AERATE THE SOIL, AND THE SOIL WILL FEED THE PLANT
The complexity of plant-soil-water-nutrient-microbial interactions is such that agronomic research which studies and varies only one practice or maybe two practices at a time – ‘controlling’ all the other variables that affect crop performance, which means excluding any consideration of their effects -- is likely to ensure that the results of such analysis will be sub-optimizing. Research findings would be understood differently if scientific reporting informed us not that variables X, Y and Z were controlled, but rather that their effects were ignored!
The phyllochron model discussed above may appear complicated, but it rests on very simple elements that have very concrete and significant implications for crop performance. It is thus worth investing some effort in understanding.
As we looked into the agronomic literature, we found various reports of relationships that were often difficult to measure but simple in their explanations, that threw light on what we were observing in the field. For example, it was reported in the literature that when the soils in rice fields are alternately flooded and dried as recommended for SRI, the nitrogen in these paddy soils will occur in both ammonium (NH₄+) and nitrate (NO₃-) forms. On the other hand, when soils are kept flooded and remain anaerobic, the N available will be mostly or only in the form of ammonium, as nitrate requires oxygen for its molecules to be synthesized.
As noted in Chapter 4, a given amount of nitrogen in the soil will support a higher rice yield if this N exists and is taken up in both ammonium and nitrate forms. This boosts yield by 40 to 70% compared to having soil N available only as ammonium, as occurs in flooded fields. Alternate wetting-drying irrigation improves rice crop performance in a number of ways, but this single effect of water management on rice plants’ nitrogen uptake is clearly non-trivial. Along similar lines, we found in the literature that biological nitrogen fixation in soils is increased by as much as 10-15 times when layers of aerobic and anaerobic soil are juxtaposed.
Several studies were done by Indian and Chinese scientists who wanted to understand how and why SRI methods result in better crop performance. One focus was on rice crops’ nitrogen-use efficiency (NUE) when cultivated with SRI vs. standard methods. In studies that applied different amounts of N fertilizer (or none) with both sets of cultivation methods, grain yield peaked at 90 kg N per ha in the Indian SRI trials and with 80 kg N per ha in the Chinese SRI trials.
Applying more fertilizer beyond these levels did not raise yield and added to costs of production. In conventional rice-growing trials in India, the optimum rate of N application was found to be 120 kg per ha, one-third higher than the optimum with SRI methods. And even when these conventionally-grown rice plants were given additional fertilizer, their yield was 30% lower than with SRI management, 4.37 tonnes per ha compared to 6.31 tonnes.
Similar trials in China showed SRI yields ranging from 5.6 to 7.3 tonnes per hectare with different levels of N fertilizer applied, ranging from zero to 240 kg per ha. With traditional transplanting and flooding, rice yield ranged from 4.1 to 6.4 tonnes with the same range of fertilizer application. With just 80 kg of N per ha, SRI plants yielded 7.08 tonnes per ha for the two seasons of trials, while conventionally-managed rice plants with 3x more fertilizer had an average yield 1 tonne lower. In China, applying 240 kg of N per ha (or more) is widespread, so there is great economic waste in rice production, as well as significant environmental degradation.
Such findings, together with SRI results in the field, lend support to the mantra of organic agriculture: “Don’t feed the plants; feed the soil, and the soil will feed the plants.” While this may sound simplistic and even dogmatic, both scientific research and practical experience support it.
To this can be added the advice that one should also actively aerate the soil, so that the benefits of having more organic matter in the soil can be more fully realized. Abundant aerobic microorganisms, fungi as well as bacteria, are needed in the soil to decompose the organic material that is available there, and these organisms all need air. Aeration of the soil can be passive, by no longer keeping rice paddies and other fields always inundated; or it can be active, for example, breaking up the soil’s surface with a mechanical weeder or some other implement. Preferably soil aeration is both active and passive.
As seen in this chapter, rather than providing plants with more nitrogen, phosphorus and other nutrients, trying to prevail upon them to grow more vigorously and productively, it will be more effective and more efficient to improve plants’ growing conditions, so that the plants will have more need for N and will therefore be more disposed to take up more of this nutrient. Optimizing the conditions for growth for plants will increase their demand for nutrients.
Plants’ need for nutrients should not be taken for granted, assuming that supplying them with N and other nutrients will give sufficient impetus and support for plant growth. Nutrients are a necessary but not a sufficient factor for increasing crop yield. To the extent that greater production can be attained from farmers’ fields with reduced provision of external, purchased nutrients – because SRI-grown plants take them up more abundantly and utilize them more effectively -- this will lower farmers’ costs of production and make agricultural production more profitable as well as less deleterious for the natural environment and for human health.
The fixation on nitrogen noted above, citing a number of eminent crop scientists, has been unfortunate. Some skeptics have balked at accepting the validity of SRI methods and results because they believed that there was not enough nitrogen in the soils or in the plants to support such profuse growth. However, as seen in this chapter the supply and uptake of nitrogen are much more complex than reflected in standard soil testing, which gives little consideration to soil biology, and in typical studies of the uptake of inorganic N.
As discussed in Chapter 11, one of the advantages of SRI crop management is its acceleration (shortening) of the crop cycle. Reducing the time required to go from seed to seed can be explained in part by the effect that transplanting young seedlings has on speeding up the plants’ biological clock. Understanding and taking advantage of Katayama’s concept of phyllochrons can help to explain this difference in rice plants’ expression of their genetic potential, yielding more productive and more robust phenotypes. But this is just one of many effects involved.
Of course, the supply of and demand for nutrients need to be both sufficient and synchronized for getting optimum results. Demand does not create its own supply any more than supply creates its own demand. But in the course of plants’ evolution, they have hit upon various mechanisms that enable their roots to attract and to interact with beneficial soil microbes, rhizobia, mycorrhizal fungi and other organisms that live symbiotically with plant roots, mutually benefiting them. For example, some microorganisms synthesize biological compounds known as siderophores that enable them and the plants associated with them to acquire iron ions from their soil environment for mutual benefit, also enhancing plants’ systemic immunity to pathogens.
This information is offered only to give an indication of how complex are the plant-soil-water- nutrient-microbial interactions that deserve consideration when trying to understand the observed and measured effects on plant growth and health that are associated with the management practices assembled under the designation of SRI. The nutrient uptake processes and plant growth patterns discussed in this chapter are not only themselves complex but they are also parts of larger complex interactions.
The various pieces of knowledge reviewed here emanate from diverse scientific sources. The most unexpected came from a Japanese plant scientist who worked in relative obscurity almost a century ago. All contributed to an emerging understanding of how the farmers around Ranomafana National Park in Madagascar, on very poor soil, with the same varieties, and without relying on purchased agricultural inputs, could quadruple their yields with healthier, climate-stress-resilient rice plants.
The ideas and observations in Chapters 4, 5 and 6 lay the scientific groundwork for an empirical but theoretically informed understanding of SRI’s effectiveness. As seen in the following chapters, there was still more work to be done. We needed to validate and calibrate the results of SRI management and to understand further benefits and extrapolations that could be attained by making changes in plants’ growing environments rather than just in their genes or by supplying external nutrients.
One needs to pay attention to genetic potentials, of course, since it is these that were being tapped by new management methods. But they were starting points rather than deterministic factors. This is considered in Chapter 10 after reviewing the research efforts undertaken by a multiplicity of scientific colleagues from many disciplines as well as thoughtful and observant practitioners.
NOTES AND REFERENCES
 .J.D. Kirk and D. R. Bouldin. ‘Speculations on the operation of the rice root system in relation to nutrient uptake,’ in F. W. T. Penning de Vries et al., eds., Simulation and Systems Analysis for Rice Production, p. 199, Wageningen: Pudoc (1991). For more detail on this relationship, see J.K. Ladha, G.J.D. Kirk, J. Bennett, S. Peng, C. K. Reddy and U. Singh, ‘Opportunities for increased nitrogen-use efficiency from improved lowland rice germplasm,’ Field Crops Research 56, pp. 46-49 (1998).
 This is often referred to as ‘chemical fertilizer’ or as ‘synthetic fertilizer.’ We use the adjective ‘inorganic’ to differentiate man-made fertilizer, such as nitrogen fertilizer made by synthesizing ammonia (NH3) from atmospheric N2 and H2, from rock phosphate that is produced from mineral deposits. All forms of fertilizer are ‘chemical’ in composition, but they will have either organic or inorganic sources.
 See table in Chapter 4, from Ana Primavesi’s Manejo Ecológico do Solo, 1st edition, Nobel, Sao Paulo, Brazil, Table 3.5, p. 94 (1980).
 N.C. Brady and R.R. Weil, The Nature and Properties of Soils, 14th edition, Prentice-Hall, Upper Saddle River, NJ (2007).
 Source given in endnote 3.
 This book was originally published in Portuguese (endnote 3) and translated into Spanish, but unfortunately not into English. Primavesi has been one of the most influential agronomists and agroecologists in Latin America, but her work is little known outside of this region.
 One of the comments in Fr. Laulanié’s unpublished technical papers which made the greatest impression on me was a statement that he regarded the rice plant as his teacher (“mon maître,” my master) from which he sought to learn, seeking to understand what would best meet its needs.
 The paper was published 11 years later: N. Uphoff, ‘Rethinking the concept of 'yield ceiling' for rice: Implications of the System of Rice Intensification (SRI) for agricultural science and practice,’ Journal of Crop and Weed, 9: 1-19 (2013).
 H.J. Kronzucker, G.J.D. Kirk, N. Y Siddiqi and A.D.M. Glass, ‘Effects of hypoxia on 13NH4+ fluxes in rice roots,’ Plant Physiology, 116, p. 581 (1998). These authors cited as support for their statement the article cited in endnote 1 by Cassman et al. (1998).
 J.K. Ladha, G.J.D. Kirk, J. Bennett, S. Peng, C.K. Redd, and U. Singh, ‘Opportunities for increased nitrogen-use efficiency from improved lowland rice germplasm. Field Crops Research 56, p. 41 (1998), citing S.K. DeDatta and R. Buresh, ‘Integrated nitrogen management in irrigated rice,’ in B.A. Stewart, ed., Advances in Soil Science, Vol. 10, 143-169, Springer, New York (1989).
 Source given in endnote 1.
 T.R. Sinclair, ‘Agronomic UFOs waste valuable scientific resources,’ Rice Today, 3: 43 (2004).
 It was interesting that while IRRI scientists were rejecting SRI as being not true or not as effective as reported, after they learned about SRI they began working on ‘alternate wetting and drying’ (AWD), one of SRI’s core practices, as a major component of their research program.
 S.B. Peng, R.J. Buresh, J.L. Huang, X.H. Zhong, Y.B. Zou, J.C. Yang, G.H. Wang, X.Y. Liu, R. Hu, Q.Y. Tang, K.H. Cui, F.S. Zhang and A. Dobermann, ‘Improving nitrogen fertilization in rice by site-specific nutrient management: A review,’ Agronomy for Sustainable Development, 30: 649-656 (2010). The authors noted that China’s N-application rate per hectare was double that of Japan’s, yet average rice yields in the two countries were similar.
 Source given in endnote 9, page 581. These two processes refer to the evaporation of N into the air in gaseous form and the microbial transformation of N into less-available forms.
 Source given in endnote 10, page 43. The IRRI work in the 1990s is reported in J.K. Ladha and P.M. Reddy, The Quest for Nitrogen Fixation in Rice, IRRI, Los Baños, Philippines (2000). By 2002, BNF had disappeared from IRRI’s research agenda, which was then devoted largely to genetic improvements and modifications in rice. BNF trials had not proven able to provide plants (directly) with as much N as could be furnished in inorganic form in bags of N fertilizer, so interest in BNF evaporated at IRRI, not appreciating adequately the role that this process and the organisms involved in it play within well-functioning, fertile soil systems.
 This is the basic concept behind the theory of ‘trophobiosis,’ proposed by Francis Chaboussou, which will be discussed in another chapter, an important contribution to the theory of SRI. Healthy Crops: A New Agricultural Revolution, Jon Anderson, Charnley, UK (2004), originally published in French in 1985. This observation is not challenged by mainstream crop science, but it gets eclipsed by the emphasis that is placed on nitrogen as the preeminent macronutrient, as expressed below.
 It was curious that Ladha et al. (1998) reported the following relationship without considering its implications for the heavy use of inorganic-N fertilizer and as a justification for paying more attention to biological nitrogen fixation. They stated that to get the best rice yields, varieties that mature more quickly need applications of 150-200 kg of N per ha, while for medium-term lines, the optimum point is about 150 kg per ha, and for long-duration varieties, it is only about 100 kg per ha.
Since scientists consider the volatilization of N (its loss into the atmosphere in gaseous form) when provided as fertilizer to be a big problem, one should expect that the longer-maturing varieties would need higher rather than lower rates of N fertilization to reach their maximum yield. These data called into question the accepted understanding of rice plants' needs and their sources of N, but the data were reported without critical examination.
 Source given in endnote 10, pages 59 and 60-61.
 Source given in endnote 1, page 199.
 Source given in endnote 1. Ladha et al. (1998) went into this matter in more detail on pages 47-48, reporting that the maximum influxes of N into the root occur only when the plant has a very low, i.e., insufficient content of N. Conversely, when the plant has higher, i.e., sufficient nitrogen status, the influx of N into the roots is suppressed by certain regulatory mechanisms. N influx is down-regulated as the plant's internal N status is increased, and it is up-regulated when the plant's N status decreases. The plant is thus interacting with its environment, seeking to meet its growth needs in an optimizing way, rather than passively accepting the nutrients that are provided to it.
 A former chief executive of the UK’s Natural Environmental Research Council, John Lawton, has characterized the overuse of N fertilizer as “the third major threat to our planet, after biodiversity loss and climate change.” Nature, 24 February, 2005. An assessment of the economic costs to countries within the European Union from their (over) use of nitrogen fertilizers has estimated that these costs add up to between 70 and 320 billion euros annually. M.A. Sutton, C.M. Howard and J.W Erisman, The European Nitrogen Assessment: Sources, Effects and Policy Perspectives. Cambridge University Press, Cambridge, UK (2011).
 McHugh’s thesis research is reported in O. McHugh, J. Barison, T.S. Steenhuis, E.C.M. Fernandes and N. Uphoff, ‘Farmer implementation of alternate wet-dry and non-flooded irrigation practices in the System of Rice Intensification (SRI), in B.A.M. Bouman H. Hengsdijk, B. Hardy, P.S. Bindraban, T.P. Tuong and J.K. Ladha, eds., Water-wise Rice Production: Proceedings of the International Workshop on Water-wise Rice Production, 8-11 April 2002, 89-102, International Rice Research Institute, Los Baños, Philippines (2002).
 The results of Barison’s thesis research were published with the author in the journal Paddy and Water Environment, ‘Rice yield and its relation to root growth and nutrient-use efficiency under SRI and conventional cultivation: An evaluation in Madagascar,’ 9: 65-78 (2011). The data that follow were reported in that article.
 SRI plants’ uptake of P was 21.03 kg per ha, and that for farmer-method plants was 12.69 kg P per ha. For potassium, the difference was more like N than P, 108.64 kg of K per ha vs. 56.77 kg K per ha.
 C. Witt, A. Dobermann, S. Abdulrachman, H.C. Gines, W. Guanghuo, R. Nagarajan, S. Satawatananont, T.T. Son, P.S. Tan, L.V. Tiem, G.C. Simbahan,and D.C. Olk, ‘Internal nutrient efficiencies of irrigated lowland rice in tropical and subtropical Asia. Field Crops Research, 63: 113-138 (1999).
 The methods used for the modeling are explained in Joeli’s master’s thesis, Nutrient-use Efficiency and Nutrient Uptake in Conventional and Intensive (SRI) Rice Cultivation Systems in Madagascar, for the Department of Crop and Soil Sciences, Cornell University, Ithaca, NY (2003).
 The solid lines in the graph below are regression lines fitted to a polynomial equation to calculate the highest coefficient of determination (R²).
 That idea that SRI results exceed some biological ‘maximum’ for rice yield has been put forward by Achim Dobermann, ‘A critical assessment of the system of rice intensification (SRI),’ Agricultural Systems, 79: 261-281 (2004); and by J.E. Sheehy, S. Peng, A. Dobermann, P.L. Mitchell, A. Ferrer, J.C. Yang, Y.B. Zou, X.H. Zhong and J.L Huang, ‘Fantastic yields in the system of rice intensification: Fact or fallacy? Field Crops Research, 88: 1-8 (2004). Their calculations follow the analysis set forth by S. Yoshida in his classic book, Fundamentals of Rice Crop Science, International Rice Research Institute, Los Baños, Philippines (1981).
 Sugarcane, for example, is a member of the grass family. A discussion of phyllochrons in sugarcane, with pictures, is given in B. Gujja, U.S. Natarajan and N. Uphoff, ‘Sustainable sugarcane initiative: A new methodology, its overview, and key challenges,’ in P. Rott, ed., Achieving Sustainable Cultivation of Sugarcane, Vol. 1, 45–76, Burleigh-Dodds, Cambridge, UK (2017). Timing and other details are different for this crop than for rice, but the processes of tillering and growth are very similar.
 T. Katayama, ‘Analytical studies of tillering in paddy rice,’ Journal of the Imperial Agricultural Experiment Station, 1: 327-374 (1931), in Japanese with English summary, and Ine mugi no bungetsu kenkyu (Studies on Tillering in Rice, Wheat and Barley). Tokyo: Yokendo Publishing (1951).
 There is a 4-page entry on ‘phyllochrons’ in the 2ⁿᵈ of the three volumes of the encyclopedia of Japanese rice science, translated into English and edited by T. Matsuo, K. Kumazawa, R. Ishii, K. Ishihara and H. Hirate, Science of the Rice Plant, Food and Agriculture Policy Research Center, Tokyo (1993). There is also a very technical discussion by K. Hanada in ‘Differentiation and development of tiller buds in rice plants, Japanese Agricultural Research Quarterly, 16: 79-86 (1982). When I visited IRRI in 1999, the only person whom I could find there who knew anything about phyllochrons was its Deputy Director-General, Dr. O. Ito, from Japan.
 K. Nemoto, S. Morita and T. Baba, ‘Shoot and root development in rice related to the phyllochron,’ Crop Science, 35: 24-29 (1995). This was the best explanation of phyllochrons that I found in English.
 ‘Le système de riziculture intensive malgache,’ Tropicultura (Brussels), 11: 110-114. The concept and its relevance to rice production was discussed in my first article on SRI published in Environment, Development and Sustainability, 1: 297-313 (1999), and elaborated on in Stoop, Uphoff and Kassam, ‘Research issues raised for the agricultural sciences by the System of Rice Intensification (SRI) from Madagascar: Opportunities for improving farming systems for resource-limited farmers,’ Agricultural Systems, 71: 249-274 (2002). The most extended discussion is in Uphoff, The System of Rice Intensification: Responses to Frequently Asked Questions (2015), pp. 154-161.
 Discussed in Nemoto et al. (1995) and reviewed below. The concept of ‘leaf age’ used in agronomy is related to the ideas of phyllochron, but it is more descriptive than explanatory. In terms of leaf age, the recommended SRI time for planting is during the 2-3 leaf stage. This may begin before 15 days after seeding in the nursery when the ambient temperature is higher, as at lower elevations and along the coast of Madagascar. It may be after 15 days in colder, higher elevations on Madagascar’s central plateau.
 I could not bring this rice plant stump back to Cornell because of government phytosanitary restrictions, but I could take this picture and give the stump to my colleague Prof. Iswandi Anas, IPB. The size of the root system is as remarkable as the number of tillers. As seen below, a rice plant completing 12 phyllochrons of growth would have 84 tillers, according to the analysis of Katayama as elaborated by Laulanié.
 Fibonacci was an Italian mathematician born in 1175 AD who among other things wrote about the characteristics of this particular sequence of numbers, now named after him (although it was discovered by an Indian mathematician about 500 years earlier). Fibonacci asked the question: what is the expansion of numbers if one has a pair of rabbits which reach sexual maturity after one month, and their gestation of a pair of offspring (male and female) takes another month, and then they start reproducing and giving birth to pairs of rabbits like themselves, and all keep reproducing in the same way indefinitely? How does the total number of rabbits increase?
It turns out that, with a one-period lag, the number produced in any period is the sum of the previous two periods: 1, 0, 1, 2, 3, 5, 8, 13, 21, etc. and the cumulative numbers are similar: 1, 1, 2, 3, 5, 8 … This semi-exponential expansion, which reflects a difference between biology (with its gestation ‘lag’) and purely mechanical production, is found fairly often in nature in terms of seed numbers and location or plant leaf branchings. An interesting discussion of this is found in B. Goodwin, How the Leopard Changed Its Spots: The Evolution of Complexity, Ch. 5, Princeton University Press, Princeton, NJ (1994). This patterning, which greatly intrigued Laulanié, is seen in the table below. Each successive period produces about 2/3 more tillers than in the preceding period, for a coefficient of about 0.65.
 To some readers, this discussion may sound teleological or anthropomorphic, and it is to some extent. The words used here are intended to sketch out for readers a logic, attributable to evolutionary success, that makes observed plant behavior coherent.
 Again, the comments made in the preceding endnote apply here. In fact, there is some small percentage of plants that have evolved more aggressive, even predatory growth strategies. But rice plants follow more of a ‘live and let live’ strategy.
 This table was worked out from reading Nemoto et al. (endnote 33) to provide an explanation of phyllochrons in my book on The System of Rice Intensification: Responses to Frequently Asked Questions, 154-161 (2015).
 H.J. Kronzucker, M.Y. Siddiqui, A.D.M. Glass, and G.J.D. Kirk, ‘Nitrate-ammonium synergism in rice: A subcellular flux analysis,’ Plant Physiology, 119:1041-1045 (1999).
 F.R. Magdoff and D R. Bouldin, ‘Nitrogen fixation in submerged soil-sand-energy material media and the aerobic-anaerobic interface,’ Plant and Soil, 33: 49-61 (1970). We also learned about something called ‘the Birch effect’ which results from wetting dry soil, triggering microbial mineralization of soil carbon and soil nitrogen, and making pulses of C and N available for plant use. H.F. Birch, ‘The effect of soil drying on humus decomposition and nitrogen, Plant and Soil, 10: 9–31 (1958).
 A.K. Thakur, S. Rath and K.G. Mandal, ‘Differential responses of system of rice intensification (SRI) and conventional flooded-rice management methods to applications of nitrogen fertilizer,’ Plant and Soil, 370: 59-71 (2013). It was found that the SRI plants at all levels of N fertilizer application, from zero to 90 kg per ha, had a higher uptake of N, greater NUE, and 49% more partial factor productivity.
These advantages were attributed to significantly higher levels of N and chlorophyll in the SRI plants’ leaves and to their greater efficiency in photosynthesis. These factors were associated with delayed leaf senescence (slower aging in SRI plants), a longer period of photosynthesis, and improved root and shoot activities. These represented phenotypical differences in the SRI plants given that there was no difference in their genotypes.
 L.M. Zhao, L.H. Wu, Y.S. Li, X.H. Lu, D.F. Zhu and N. Uphoff, ‘Influence of the System of Rice Intensification on rice yield and water and nitrogen use efficiency with different application rates,’ Experimental Agriculture, 45: 275-286 (2009). They found that the highest N applications reduced yield with both SRI and conventional methods, and also lowered NUE and partial factor productivity.
Already 15 years ago, the levels of nitrite in groundwater supply in many parts of China exceeded 300 parts per million, 6x the level acceptable to the US Environmental Protection Agency, according to a paper by J.L. Hatfield and John Prueger, USDA, presented at the 4th International Crop Science Congress in Brisbane, Australia: ‘Nitrogen over-use, under-use and efficiency’ (2004). Applications of N fertilizer per ha at the time ranged often from 500 to 1,900 kg per annum.
 This idea is often traced back to Sir Albert Howard, whose book An Agriculturalist’s Testament, Oxford University Press, Oxford, UK (1943), helped to launch what became known as ‘organic farming’ in India, Great Britain and beyond. The Rodale Institute which has given leadership for organic farming in the US has similar advise: Feed the Soil: Rodale’s Complete Guide to Soil Improvement, Kutztown, PA (1995). See also the Rodale blog by P. Hepperly, ‘Feed the soil, not the plant,’ and E. Back, Feed the Soil, Not the Plants: The Organic Gardener’s Mantra (2012).
 See endnotes 9 and 10.
 A colleague and renowned expert on both rice and tropical soils, Pedro Sanchez, for example, several times (in 1999 and 2004) told me that he could not accept SRI claims until it was shown to him ‘where the nitrogen comes from.’ Pointing out that the N obviously had to come from somewhere to support the large visible increases in growth and yield did not satisfy him.
 It is interesting that the person most responsible for agronomists’ preoccupation with nitrogen, the German soil scientist Justus von Liebig, eventually recanted the primacy that he had assigned to this macronutrient.
Von Liebig’s famous ‘law of the minimum,’ articulated in 1840, argued that there will always be some nutritionally-limiting factor not sufficiently available in the soil which will constrain crop growth. He emphasized nitrogen as the plant nutrient that is most often constraining.
By the end of his career, however, von Liebig took a more holistic view of soil systems. He no longer emphasized single elements and paid more attention to living components. Reflecting on his life’s work, he wrote in 1865:
“In the years 1840 to 1842, I proposed that the natural sources which deliver to plants the nitrogen they need are not sufficient for the [production] objectives of agriculture. A series of observations as well as continuous reconsideration have indicated to me, however, that this view is not correct… For millennia, millions of people have believed, and millions believe it still, that the sun revolves around the earth because this is what they perceive.
“In the same way, many thousands of farmers have believed, and thousands still believe, that the practice of agriculture revolves around nitrogen, even though this belief has never been scientifically validated, and never will be scientifically supported because all progress and indeed all improvements in agriculture revolve around the soil” (author’s translation from German).
‘Der Stickstoff im Haushalt der Nature und in der Landwirtschaft’ (Nitrogen in the realm of nature and in farming), in Naturegesetz im Landbau: Es ist ja dies die Spitze meines Lebens (Natural Laws in Agriculture: Essays at the Culmination of My Life), 12-13, Stiftung Ökologie Landbau, Bad Dürkheim, Germany (1995).
This rethinking by von Liebig was called to my attention by an SRI colleague in Cambodia, Dr. Yang Saing Koma, an agronomist who championed the introduction and spread of SRI in his country. Koma had done his PhD studies in Chemnitz, Germany (at a time when it was still known as Karl-Marx Stadt), and he thus knew German well. This disclaimer by von Liebig is little known among agronomists who cannot read German as Koma could. This recantation is also not well-known among German agronomists.
An additional revisionist consideration is that for decades, agronomists have maintained that plants take up nitrogen from the soil only in inorganic forms (NO3 or NH4). But research in the last decade has shown that plants also take up N in organic forms, for example, T. Näsholm, K. Kielland and U. Ganetag, ‘Uptake of organic N by plants,’ New Phytologist 182: 31-48 (2009). This finding diminishes the requirements and benefits claimed for applying, especially large amounts of, inorganic N fertilizer.
 D.V. Badri and J.M. Vivanco, ‘Regulation and functions of root exudates,’ Plant, Cell and Environment, 32: 666-681 (2009).
 J.B. Nieland, ‘Siderophores: Structure and functions of microbial iron transport compounds,’ Journal of Biological Chemistry, 270: 20723-20726 (1995); A. Aznar and A. Dellagi, ‘New insights into the role of siderophores as triggers of plant immunity,’ Journal of Experimental Botany, 66: 3001-3010 (2015).
PICTURE CREDIT: Norman Uphoff (Cornell)