Abstract
Basic science aims at understanding nature. Because scientific‘understanding’ is supposed to be free of the opinions andexperiences of the individual investigator, the most importantcriterion for the quality of empirical studies is the independencyof the results of the investigator, that is, their‘objectivity’ or,more correctly, their ‘inter-subjectivity’. Whether this goal hasbeen achieved usually is measured in terms of reproducibility:ideally, a researcher B who repeats an experiment performed byresearcher A should observe the same results and draw the sameconclusions. Surprisingly, the central goal of science and thisquality criterion often result contradictory to each other, partic-ularly in ecology. Ecological interactions are context-dependent,which enhances the variability in the data and seems to dramat-ically reduce their reproducibility. The common solution is tomove from the field to the greenhouse or growth chamber andwork with genetically homogeneous populations, to control abi-otic and biotic factors and avoid noise that results from usingdifferent genotypes.Unfortunately, by aiming at maximizing the reproducibility ofour results we frequently lose their ecological and evolutionaryrelevance. Natural populationsare genetically variable. Thisvariation represents an important source of adaptive, evolution-ary changes, because selection can act only on variable traits.Furthermore, organisms are phenotypically plastic, and the ca-pacity to express different phenotypes when facing differentenvironments represents a crucial fitness-enhancing trait. Thus,variation that we perceive as ‘noise’ in our data might ratherindicate a vitally important plasticity of the organism, and study-ing an organism under one specific set of environmental condi-tions might reveal just one single out of thousands of possiblephenotypes, and perhaps a phenotype that is unlikely to occur innature at all. How can we get rid of the noise without losingecological relevance?Of course, the above-mentioned lack of reproducibility ap-plies only as long as we do not understand all factors thatinfluence our process of interest. As soon as we understand allfactors of relevance, we will obtain reproducible data. Therefore,I argue that (i) a zig-zag course between field and laboratory islikely to provide us with relevant AND reproducible data forecological research and that, (ii)‘noise’ is not necessarily some-thing we simply have to get rid of,but rather a putative indicatorof interesting phenomena that we do not yet understand.Perhaps, the most famous scientific breakthrough that was initi-ated by a failure in controlling the experimental conditions is thediscovery of antibiotics. If Alexander Fleming had nicely con-trolled his experimental conditions and avoided the unforeseeninfection of his cultures of Staphylococci by an environmentalPenicillium, he never would have discovered the antibiotics.Rather, Fleming worked under non-sterile conditions, made amistake, carefully observed the derivation from the expectedoutcome of the experiment, profoundly thought about possibleexplanations, formulated a hypothesis aimed at explaining thisphenomenon, tested the hypothesis, observed reproduciblechanges in bacterial growth rates and, thereby, discoveredantibiotics!Similarly, although admittedly of much lower importance andrequiring way lower levels of creativity, the discovery thatextrafloral nectar (EFN) secretion byMacaranga tanariusrepre-sents a jasmonic-acid (JA)–mediated induced defense trait startedfrom a noisy dataset. A field study aimed at quantifying themetabolic cost of EFN suffered from huge levels of variation.Even organizing the data by leaf age did not help much, althoughit became clear that leaf ontogeny affects EFN secretion.However, then I observed a positive correlation of EFN secretionrates with standing levels of leaf damage. Of course, correlationsdo not mean too much. In fact, a positive correlation of a planttrait with standing levels of damage can indicate two contrastingphenomena: induced resistance (when the damage comes first andthen causes stronger expression of the trait, which consecutivelyhindersfurtherdamage) or inducedsusceptibility (when thestron-ger expression of the trait comes first and then enhances damage).Anyway, the multiple reports on JA-dependent induced resistancetraits in other plants made it easy to formulate a hypothesis andtest it in the field. Then, knowing that herbivory and mechanicaldamage indeed do induce EFN flow, it was time to go to the laband work on the underlying mechanism (Heil et al.2001)ofaprocess that by this time was known to be of relevance for thisplant in its natural habitat. Similarly, Rhoades in the 1980s first
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.