The study of biology in general, and evolution or ecology in particular, may be distinguished categorically in a number of ways. For example, analyses may or may not employ time as a variable. Studies also can be observational versus experimental, or take place in situ versus ex situ versus in vitro. It is also possible, and helpful, to distinguish among studies that are inherently more ecological versus those which are more evolutionary biological. Microbial experimental evolution studies, for example, typically employ a time variable, and more often than not take place within artificial, i.e., in vitro environments. Whether or not such analyses are more observational versus experimental, however, is dependent, as much as anything, on the degree to which the "experiment" is controlled in terms of comparison between "treated" versus "untreated" runs. Below I provide greater elaboration on such thinking.
Observational, experimental, or theoretical approaches may be more static than dynamic. A static analysis is a snapshot, or multiple snapshots, of one or multiple subjects taken at one time, or at least at some approximation of a single point in time. Taking measurements along transects, for example, is static, at least in the sense that ideally all such measurements would be made simultaneously. Sequencing the genome of an organism is static, also, in the sense that what is being observed is the organism as it exists at a single moment, though ideally a moment that captures, as well as possible, the condition that existed in the wild, pre-isolation state, i.e., pre-laboratory domestication in the case of microorganisms. Similarly, individual metagenomic analyses can be described as snapshots of the genetic structure of a community that existed at a single moment in time. Of course, explicitly one may take similar measurements of different subjects, e.g., the genomic sequences of multiple isolates, the metagenomes of different places, or phage densities on mountain tops versus the bottoms of deep sea trenches. Nevertheless, time, with static analyses, either principally or ideally, is not a variable.
Analyses alternatively can take place in which time is very much a variable. Here many of the same techniques may be applied as in a static analysis, except that measurements explicitly are made in series so that a given present state may be compared with specific past or future states, all of the same system, or at least some reasonable approximation of the same system. We can further classify such temporally varied or longitudinal studies into ones in which the system is versus is not experimentally perturbed. Much of experimental biology has involved experiments in which some action is performed on a system, or equivalent parallel system, with measurements minimally taken of its state both prior to the performance of the specific perturbation and sometime after.
Alternatively, a more observational approach to biology is to perform measurements on a system that has not been intentionally perturbed, and ideally is minimally affected by what is involved in taking measurements. Perhaps especially to the sciences of evolutionary biology and ecology, experimental approaches that involve intentional perturbation are no more or less relevant than more observational approaches that do not involve perturbations. Instead, experimental approaches typically are easier, or at least less time consuming, and consequently are at least potentially more powerful, but at the same time are inherently artificial. One thus can follow a real ecosystem over one or more seasons and thereby (one hopes) obtain a fundamental understanding of the dynamics of a given system, or one can artificially create a subset of such a system, a model, which is then manipulated experimentally or theoretically in a specific, ideally well-controlled manner. Of course, perturbations not only involve a time variable, either explicitly or implicitly, but a perturbation also may vary, qualitatively or quantitatively, in terms of the actual perturbation. For example, one could remove all of the virus particles from a microcosm suspended within a pond and then in some manner measure what happens. Alternatively, one could add viruses, or provide shade versus varying degrees of sunshine, or artificial lighting, etc.
In terms of temporal analyses, one additionally can consider ecological, evolutionary, or even evolutionary ecological investigations. As above, these approaches may be made with versus without experimental perturbation. Also as above, while with-perturbation may be more powerful, that power inevitably comes at a cost of reduced realism. Nevertheless, an ecological analysis, ideally, does not involve evolutionary change in the subjects involved. Furthermore, again ideally, within-population genetic differences between individuals, ones that result in Darwinian fitness differences over the course of the experiment, are not crucial aspects of the basic design of the ecological experiments. Alternatively, in an evolutionary ecological experimental or observational study, genetic differences that give rise to fitness differences are a key, indeed, likely the key aspect of analysis. Evolutionary ecology in other words is the study of the impact of adaptations on organismal fitness. Such analyses are designed specifically to measure how genotype frequencies vary over time. Analyses are controlled, however, in the sense that at a minimum the basic fitness characteristics of the dominant players are known, e.g., various microorganisms. Furthermore, from the perspective of microbiologists, genotypic as well as phenotypic differences among these players ideally are understood more robustly than simply in terms of their impact on organismal fitness.
With regard to changes in genotype, evolutionary studies can be less-well controlled than evolutionary ecological ones. Here not only can the impact of natural selection on genotype frequency play important roles in the outcome of the analysis, but so too can genetic drift, migration, and mutation. Indeed, a typical evolutionary experiment involves exposing a population to specific environmental conditions and then passaging the population through time in large numbers such that mutations accumulate and also so that mutations, as novel genotypes, can compete. In the end, the goal often is to characterize accumulated variants and, from this characterization, make inferences about the evolutionary, ecological, or evolutionary ecological dynamics associated with the population or its environment. Such analyses can involve sequencing or otherwise phenotypically characterizing subsets of natural populations, including over time, and then to make retrospective inferences with regard to evolutionary, ecological, or evolutionary ecological dynamics within those populations. From these varied approaches one can gain robust understanding of how organisms interact with their environments and how those interactions can change as a function of time.