The book entitled The Structure of Scientific Revolutions by Thomas Khun discusses how normal science proceeds and how seminal discoveries occur . The discovery of photo-enhanced toxicity of polycyclic aromatic hydrocarbons (PAHs) was an archetypical paradigm shift as described by Khun. It is a story of being in the right place at the right time; but even more, it was the vision of the U.S. Environmental Protection Agency's (U.S. EPA) to recognize the opportunity and to have the flexibility to allow the researchers to follow up on a unique observation. It is also the story of some young students' willingness take a chance and follow up on an observation, their readiness to develop the methodologies, and their ability to conduct a number of difficult assays that resulted in the comprehensive understanding and accurate predictive models that were presented in what would become one of the top 100 most cited articles in Environmental Toxicology and Chemistry . Finally, it is the story of being able to pull together information in disciplines ranging from physical chemistry, to bioaccumulation, to toxicology, to quantitative structure–activity relationship simulation modeling to develop an overall predictive model. It was only when all of these elements came together that the seminal advancement was possible.
The phenomenon of photo-enhanced toxicity has now become “normal science,” with papers routinely published on the topic and sessions at national and international meetings. Sometimes entire meetings are organized around this topic, and there are entire journals devoted to the phenomenon. At the 25th Annual Dioxin Meeting held in Toronto, Canada, in 2005, photo-enhanced toxicity of PAHs to aquatic organisms  was listed as one of the 25 most influential discoveries of the preceding 25 years (http://www.dioxin20xx.org/pdfs/history/Dioxin2005.pdf). Photo-enhanced toxicity has now been included in the development of water and sediment quality criteria.
Like many of the most interesting discoveries in science, the discovery of photo-enhanced toxicity was the result of the unexpected rather than that of a specific design. We were conducting a large mesocosm study at the University of Georgia as part of a program to develop an environmental fate model for organic chemicals (FOAM: Fate of Organic Molecules) [3, 4]. It had been decided to focus on PAHs as model compounds because it was the time of the oil embargo of the 1970s and the U.S. Department of Energy and the U.S. EPA were involved in a Synthetic Fuels Program. The 100-m–long mesocosms had been constructed and colonized over a three-year period. The system was fitted with a number of monitoring probes, and then the model chemical anthracene was added to the system at about 6:00 AM. After being up all night preparing for the initiation of the experiment, the principal investigator went back to catch a few “Z's” on the floor of his office, He had left instructions to call him if any problems arose. He had only just arrived in his office when he received a call that all of the fish had begun dying within 20 min of adding the anthracene. The study was designed as a “fates” study, and the concentrations used—which were based on laboratory studies—should not have caused any toxicity.
While at first bemoaning the fact that the huge experiment was ruined, the investigators soon became intrigued by this unexpected result and were determined to discover what was causing it. At first, it was thought the effect could be due to the carrier solvent (ethanol) or some interaction between the solvent and the PAHs; but the patterns of where and when toxicity was observed and which organisms seemed to be more sensitive gave rise to a working hypothesis of an interaction between sunlight and anthracene. The most obvious explanation was a photochemical reaction, and the focus settled on the potential toxicity of reaction products of photolysis; but the obvious product, which was benzoic acid, was known to be insufficiently toxic to cause the observed lethality. After a call to the U.S. EPA program officer, H. Holm, in the Athens Laboratory (Athens, GA, USA), the University of Georgia team was green-lighted to “work out what was causing the unexpected toxicity.” The toxicity had occurred first at the upstream end of the systems, and it was postulated that a transformation product of anthracene was the cause. Within a day, a study using the six replicate channels with sections covered upstream or downstream allowed us to determine that it was not a degradation product but rather the parent anthracene that was causing the toxicity. The effect was dramatic, with lethality in full sunlight occurring within minutes.
Further studies were designed and methods developed to monitor for incident solar radiation, and eventually it was determined that the toxicity was a function of the amount of anthracene in the organism and exposure to solar radiation [5, 6]. It was further determined, through use of selective filters, that ultraviolet (UV) radiation in the UV-A region with a wavelength range between 320 and 400 nm was causing the effect. It so happens that anthracene has a “red-shifted” absorption spectrum that allows the molecule to absorb light at wavelengths greater than the 284-nm cut-off of the atmosphere. Subsequent studies with other organisms including daphnids [7, 8], fish [9-11], and algae [12-15] supported the original study's findings that solar radiation enhanced the toxicity of anthracene. The phenomenon was then shown to occur under field conditions . It was demonstrated with a range of PAHs that the phenomenon could be described by use of physical models based on first principles of photochemistry and that the mechanism of effect was a singlet-oxygen–mediated oxidative stress that disrupted membranes [2, 16-18]. The degree of toxicity could be described as a function of concentration of PAH and intensity of solar irradiance. Several PAHs were found to exhibit the phenomenon, and the model was further refined to predict toxicity based on the octanol–water partition coefficient (KOW) that predicted how much of the chemical would accumulate, the molar absorptivity of the individual compounds, and the probability of forming singlet oxygen that could initiate production of free radicals, which could be predicted from the singlet–triplet splitting energies of individual compounds [2, 10]. Finally, a predictive model based solely on theoretical first principles related to the structure of the individual molecules was developed that had 100% correct predictive power .
We were all very excited about these results, but convincing the scientific community that this phenomenon had any environmental relevance was a hard sell because most people had been taught that UV light does not penetrate to significant depths in aquatic systems. We then took our experiments to the Laurentian Great Lakes  and Lake Tahoe [20, 21]. Lake Tahoe is very transparent to UV light  and is also at an elevation that would result in greater amounts of UV radiation. Against much resistance, we were eventually able to demonstrate that the phenomenon does in fact occur, and it occurs at environmentally relevant concentrations of PAHs and UV radiation. The toxicity of some PAHs under natural conditions of solar irradiance was 50,000-fold more toxic than under laboratory conditions in the absence of UV light. In tests with daphnids, median times to effect were on the order of seconds. In fact, at one point early in the work at the University of Georgia, when we were having trouble reproducing results, we found that the lab technician had been carrying an uncovered tray of daphnids in cups containing PAHs outside from one trailer to another. The 30 s in full sunlight was long enough to cause lethality.
It is our belief that dogma prevented the discovery of the phenomenon in the natural environment sooner. Polycyclic aromatic hydrocarbons have been known photosensitizers in humans for 100s if not 1,000s of years . Anecdotal information indicates that PAH photo-induced toxicity was observed in fish as early as the 1950s, but the cause of toxicity was not explored at the time of the report. In fact, at the time of our discovery, scientists working with PAHs were conducting studies under yellow lights to avoid photodegradation during incubation. This practice as much as anything prevented the discovery from being made sooner. The incorporation of field tests of laboratory predictions; strong foundations in biology, chemistry, and physics; and a team of scientists who were willing to explore unanticipated and seemingly negative results enabled the discovery.
This was very much a situation of serendipity ; but the moral of the story is to expect the unexpected and then be willing to set aside preconceived notions and, yes, the original intent of the study and original experimental design to follow up on unexpected but interesting observations. Never stop observing and questioning and do not be afraid to follow up on things that do not go the way they were designed. Great discoveries  are made by people who are in the right place at the right time and who have a prepared mind and a flexible approach to research that allows them to surmount dogma and the resistance of “normal science.”