Imagine an animal biologist that would have to do without the experimental approaches worked out in fruitflies and C. elegans worms, and without the resulting insights. How many factors relevant for signal transduction and development would have been identified? How clear would it be what the counterparts of oncogenes actually do in the cell? How much would be known about molecular causes underlying various diseases? How well would one understand the "logic of the genes" in developing multicellular animals? How well could the sequence of the human genome be interpreted and used to address biological function?

While thinking of these questions, it is not difficult to realize why many plant scientists have enthusiastically embraced the concept of research on genetic model organisms, as illustrated by the rapid rise of the inconspicuous little weed Arabidopsis. The recent elucidation of the complete Arabidopsis genome sequence and genomic/bioinformatic approaches accelerate the analysis of gene functions, as illustrated in the previous Plant Physiology special issue on Arabidopsis. It is certainly true that genome science involves large-scale projects to extract broad biological information from the sequence. This special issue will be published while the first cycle of grants in the Arabidopsis 2010 Project funded by the U.S. National Science Foundation compete for funding. The goal of the 2010 Project is to understand the biological function of every gene in Arabidopsis by the end of the decade using a systems approach for the efficient characterization of gene function (Chory et al., 2000; www.Arabidopsis.org/workshop1.html). In yeast and C. elegans, systematic reverse genetics surveys have already been successful (Ross-MacDonald et al., 1999; Fraser et al., 2000; Gonczy et al., 2000) and they provide examples of the types of discoveries that can be made.

Many see the genomics revolution as heralding a completely different way of doing science that will force "lab-based researchers" to adapt or die (Butler, 2001). So, is this the end of traditional physiology and biochemistry? Far from that, we are confident that it is the beginning of a new appreciation of all existing approaches to achieve the fullest possible understanding of a living organism in the plant kingdom. This possibility to "plug and play" and the gratification associated with it is no doubt a major binding factor within the Arabidopsis community. However, there are already many examples, in yeast and C. elegans as well as in Arabidopsis, to demonstrate that focused hard work on a single phenotype or gene is the road to salvation more than 50% of the time. As just one example of the relative perils of survey anal ysis, we may consider the systematic screen for insertion mutations in 73 genes encoding members of the R2R3 MYB family of transcription factors (Meissner et al., 1999). This large project (26 authors on the paper!) identified at least one mutant allele for 36 of the 73 genes. However, when 26 of these lines were subjected to 28 different growth and treatment regimes, only six lines showed any discernible phenotype. It is presumable that many of the genes encode partially redundant functions or act in pathways for which appropriate phenotypic tests were not performed.

Such cautionary tales notwithstanding, these survey projects are, of course, invaluable new tools in biology even when more detailed investigations are needed to complete our knowledge. The genomic approaches are even more important at another level of analysisthey provide a description (which may appear cryptic right now) of the boundaries of the current biological universe. The genome of Arabidopsis is the totality of genes and functions for this organism. Every phenotype we observe, every biochemical interconversion, must be explained within the context of these 26,000 elements. The articles in this second Plant Physiology special issue attest to these concepts by illustrating a steady progress of Arabidopsis research. Even from this single issue it becomes clear how investigators use many different toolsfrom classical ultrastructural analysis (Park and Twell, 2001) to state-of-the-art proteomics (Gallardo et al., 2001), depending on the question being addressed. It is obvious that we can look forward to many further discoveries being presented in the pages of Plant Physiology. For this reason, an Arabidopsis special issue is already being planned for June 2002. The editor will be Fred Ausubel.

It has been many years since recombinant cDNA expression in E. coli or eukaryotic hosts, together with facile purification techniques (through His tags, etc.), freed us from big-bucket enzymology and allowed kinetic and structural studies from even the diminutive weed. A recent example is the description, in this issue, of the three-dimensional structure of cystathione -lyase (Breitinger et al., 2001). Now, it is the development of improved techniques and equipment for chemical analyses that is allowing the Arabidopsis train to continue to accelerate. On the one hand, metabolic profiling is promising to provide a new way to quiz the organism about the consequences of specific treatments or mutations. On the other hand, better protein separation techniques and impressive improvements in the sensitivity of mass spectrometers are allowing more penetrating analysis of the Arabidopsis proteome and its subcomponents. You can read more about these new advances in the Update by Klaas van Wijk in this issue (van Wijk, 2001). It is extremely important to have the genome sequence as a benchmark resource for proteomics projects, so Arabidopsis once again finds herself in a position to disclose the mysteries of nature. Of course, "le hasard favorise l'esprit prepare."

One very tangible and comforting outcome of the full genome is the knowledge that we know exactly what we are up against as research investigators. For example, the cloning of the first cellulose synthases (Pear et al., 1996; Arioli et al., 1998) introduced us to these engaging molecular spinnerets, but understanding them fully in the context of the whole plant has been a challenge. No wonder! Arabidopsis has 10 CesA genesthe best candidates to encode the catalytic subunits of cellulose synthases. In the CesA plus Csl (Ces like) superfamily, there are a total of 39 genes which, in all probability, encode processive glycosyltransferases responsible for the synthesis of a range of polysaccharide polymers (Richmond and Somerville, 2000). A paper describing one of these, AtCSLD3, is included in this issue (Wang et al., 2001). For this family (and many others defined by the genome sequence), the task ahead is still great, but now it is finite.

Understanding the development from a single cell to a complete organism remains a daunting task, even in the animal model systems where multiple gene functions have been linked to specific steps in development. Despite the seemingly more simple body plan, a deep understanding of plant development has not been less difficult to achieve. One reason for limited progress in understanding the genetic instructions that lead from zygote to mature plant, and from gametophyte to zygote, may be the intense interplay between developmental programs and the environment. Besides obstructing easy and straightforward genetic searches for developmentally relevant genes, such interplay presupposes complicatedly wired regulatory networks.

Dissection of developmental controls has been the major objective of Arabidopsis developmental research, starting out with flower development now 15 years ago and spreading to the development of vegetative meristems, embryogenesis, and the patterning of conspicuous single cell types. An impressive list of factors involved in primordium or cell fate specification and in some cases information of their link to environmental inputs has resulted. The genome sequence will speed up this dissection considerably because it allows more rapid map based cloning of genes detected by the ever-valuable chemical or physical mutagens. Even high-throughput methods to categorize ethylmethanesulfonate-induced mutations become feasible nowadays (Colbert et al., 2001). The genome sequence also allows the dissection of redundant functions by candidate gene approaches. The now widely accepted strategy for genetic dissection used by developmental biologistsanalyzing gene function by phenotyping both loss-of-function and gain-of-function allelesis becoming common practice in the plant field. Even if loss-of-function reveals no phenotype (for example, see Baima et al., 2001), communication of that fact through publication is important for development of the larger picture. This reductionist approach will provide us with the nodes of the regulatory network that enables plant development.

An important first step toward integrating data on single genes is achieved by genetic approaches, as the combination of mutants (or transgenes) reveals connections between gene functions. Good examples are the analysis of combinations of flowering time mutants and the analysis of UV-B radiation effects (Boccalandro et al., 2001; Reeves and Coupland, 2001). Protein interaction data, which can be analyzed more systematically aided by the genome sequence, will help to strengthen these connections and to discover new "links between the nodes."

However, dissection and the analysis of local interactions, which can be equaled to the identification of the nuts and bolts of the developmental machinery, is not likely to be sufficient for understanding plant development at a deep level. The complexity of gene regulatory networks, and their wiring into basic cellular processes and environmental effects asks for integration of data at several levels. Very careful analysis of many Arabidopsis genotypes will be necessary to gain insight into integrated control of development. Quantitative effects of genetic and other alterations will have to be measured (for example, see Leyser, 2001) and modeled. Physiological data sets on mutants, DNA array analysis on mutants combined with treatments, and metabolic profiling are just three examples of the kind of quantitative data that will become more important. We expect that this will lead to a renewed appreciation of many of the refined techniques for quantitative analysis that have been developed by plant physiologists. A new niche is emerging: "developmental physiology." Plant Physiology should be an excellent podium for this niche.