Wikipedia:Reference desk/Archives/Science/2022 November 27

Welcome to the Wikipedia Science Reference Desk Archives
The page you are currently viewing is a transcluded archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

November 27

When UV breaks bonds, are majority time they are just rebonded?

For O2, it takes <241 nm of radiation to break O=O bond, that 241 nm is in the UV-C radiation. For triple bonds like N2, that requires <127 nm. Now I'm looking at plants, and C-C bond, which <343 nm of radiation can break it. That is in the UV-A range. But you don't see plants or animals being broken down by sunlight. My guess is, are these bonds instantly rebonded after a split? So if sunlight breaks an O2 bond, do the 2 singlet oxygen just rebond right back? Then is this also done in longer carbon chains like plants and animals? Maybe for complex molecules like plant/animal structures, there is rearrangements when C-C bonds are broke like 1% of the time. 67.165.185.178 (talk) 10:30, 27 November 2022 (UTC).

The solar UV radiation shorter than about 300 nm is nearly completely absorbed by the ozone layer. Ruslik_Zero 19:36, 27 November 2022 (UTC)

Really interesting question, and animals and plants do indeed break down under UV light! The reason why they appear to not is that they have a myriad of natural systems in place to block UV damage in the first place, and more systems to repair UV damage after it occurs. For example, in plants their cuticles serve to absorb UV light in the B/A range via vacuole-bound flavinoids and phenolic compounds in cell walls or cutin. Also, in most organisms DNA is susceptible to damage caused by UV irradiation, a common form being pyrimidine dimers. These are restored by dedicated protein repair mechanisms.

However, I am no physical chemist so I can't speak to what percentage of UV-induced homolytic cleavage of O2, N2, C-C bonds recombine under natural conditions. I'm sure it occurs some of the time, but not often enough for living organisms to rely on! Synpath (talk) 21:01, 28 November 2022 (UTC)

What are leaves look like on the molecular level, are they like polymers? -C-C- bonds, or? And yes, just because <343 nm can break a C-C bond, is the maximum largest nm that can. But that could mean that <330 and <320 are still very unlikely to break a C-C bond, so, I never got into the details of that. And would it not have to be at a very precise location? Sunlight is 53-55% IR, but I never found the demographics for what % is visible light and UV. 67.165.185.178 (talk) 00:19, 29 November 2022 (UTC).

So, just because a particular wavelength of light can break certain bonds, doesn't mean that every bond in every molecule will be broken instantly upon exposure to UV light. On a per-molecule basis, it is still a relatively low-probability event. On the unrelated question of what "leaves" look like on a molecular level, mostly what they look like is cellulose, which is a polymer that contains many C-C bonds; and exposure to UV light does not cause leaves to vaporize! --Jayron32 13:15, 29 November 2022 (UTC)

What generates what distribution?

Intuitively it's clear, why throwing a handful of dice and adding the result, would output normal distribution. Combinations like 6-1, 4-3, 5-2 will populate the middle, around an average of 3.5, and extreme combinations like 6-6-... or 1-1-... will be less common, and populate the left and right tail.

However, according to Pareto principle, "natural phenomena distribute according to power law statistics", for example, "80% of sales come from 20% of clients." What force would generate this? Bumptump (talk) 19:20, 27 November 2022 (UTC)

Power law is not the normal distribution. Ruslik_Zero 19:32, 27 November 2022 (UTC)

* The question is what generates what distribution, Pareto distribution, normal distribution, whatever. Bumptump (talk) 21:13, 27 November 2022 (UTC)

Purists would say that real world examples may sometimes resemble certain mathematical functions. All that is required is that the area under the function must be 1. Any attempts to assign a hypothesis or principle to a distribution or pattern of distributions is to be viewed through the lens of statistical hypothesis testing. Abductive (reasoning) 21:00, 27 November 2022 (UTC)

You missed out 'Many' in your quote. It isn't proper quoting when you leave out words like that so the 'quote' says something quite different. Power law distributions typically arise when there is a multiplicative effect in what is being measured. For instance one might expect something like it in how many people access websites or how much money people have. A popular website is talked about more and gets more resources and can develop. A rich person is more easily able to pay for better investment and tax advice and get better rates on loans. You don't get such an effect with for intance the heights of people which is better modelled with the normal distribution. NadVolum (talk) 21:31, 27 November 2022 (UTC)

My bad. Indeed the "many" in the quote changes the meaning substantially. A classical case of Quoting out of context, albeit unintended. Bumptump (talk) 21:59, 27 November 2022 (UTC)

Additionally, a power law fitting these observed "natural" distributions is merely an empirical fact without a generic underlying theoretical explanation. The fit is often mainly adequate for the tail only and even then far from perfect, saved only by the bell paucity of high-valued observations. A log-normal distribution may work as well or better. In fact, given an empirically obtained distribution of a random variable $X$ assuming values in the set of (theoretically unbounded) positive real values $(0,\infty ),$ when seeking a fit with a model distribution it is often wise to seek a fit of the empirical distribution of $\log X$ instead.  --Lambiam 14:03, 28 November 2022 (UTC)

• So, the deal is, there are many different phenomena modeled by many different mathematical functions. Some fit a power law, some fit a normal distribution, some fit other mathematical models, such as the Logistic function, the Logistic map (Which is different from the Logistic function), the Maxwell–Boltzmann distribution, etc. Some may follow a polynomial function, of which there are many, from linear, second order, third order, etc. You seem to have this sort of thing a bit backwards; first we make an empirical observation about some phenomena over time, then we find a mathematical function that matches, or models, that behavior, then we test the mathematical function against new data to see if it works in a predictive manner. That kind of back-and-forth between making observations, developing a theory that matches the observations, then testing and refining the theory based on how well it matches, is the core principle of the scientific method. Nature does what nature does, we as people create mathematical models to allow us to better predict what nature will do. --Jayron32 13:02, 29 November 2022 (UTC)

When you (@Jayron32:) say, I "have this sort of thing a bit backwards": pls, correct me if I'm wrong, but I assume that when we observe a distribution of any real world phenomena, we can at least speculate about what process would have caused it. And equally, when we understand a process, we can at least speculate about the resulting distribution. We can start looking at the distribution or at the process. For example, we can look at population growth of bacteria or at the function generating it (get a bacterium, add glucose and wait ~10 min, then you have 2 bacteria, an exponential growth).

Whether we are discovering math in the world or we are just inventing a math explanation to model an observation is a totally different question and a philosophical one. I'm interested in matching a function or some functions to a distribution and vice versa.

So, I'm puzzled by distributions that appear to be created out of thin air. For example, in the Zipf's law you pick a group of cities, the largest city is roughly twice the size of the second largest city, three times the size of the third largest city, and so on. And it's not as if some government would have decided that it should be so. The frequency of vocabulary appears to exhibit the same distribution.

At least speculatively, shouldn't we be able to propose a function (some process in society) that generates these distributions? Bumptump (talk) 17:19, 30 November 2022 (UTC)

So, the process doesn't "create" the distribution. We, as humans, create a mathematical function that generates numbers where THOSE numbers substantially match the data we get. The mathematical function is a human invention designed to be useful to us in helping us predict the date we might get from further observations. But nature does what it will, the mathematical results are stuff humans create to help us understand and explain and model the data. The "function" in this case is not a natural process, it's a man made model (simplified tool) used to predict and explain the data. --Jayron32 18:10, 30 November 2022 (UTC)

We can make mathematical models of processes, such as the repeated throwing of a triple of dice, using abstractions such as fairness of the dice and successive throws being independent and identically distributed. From such a model we can calculate the distribution of possible outcomes, including its central tendency, which is itself a (discrete) distribution. Conversely, given an observed distribution, such as a Zipfian one, we can try to find a model process giving rise to approximately that distribution. Doing so is essentially applying the general effort of constructing mathematical models fitting observed phenomena as a possible theoretical explanation – an effort without which modern science would not exist. As to Zipf's law, many theoretical explanations of its appearance in language have been proposed.^[1] Likewise, there are proposed explanations for its applicability to the population sizes of cities across different economic and cultural conditions.^[2]  --Lambiam 22:17, 2 December 2022 (UTC)

Should my mom get Tdap?

request for medical advice --Jayron32 12:51, 29 November 2022 (UTC)
The following discussion has been closed. Please do not modify it.
Yesterday, my mom took a garbage bag of rusty tack strips out, and she got hurt. But it was just a very small scratch that didn't bleed at all and didn't have a lot of dirt in it. Should my mom get a tetanus shot or no? And the last time she got a tetanus shot was 14 years ago. 67.215.28.226 (talk) 22:50, 27 November 2022 (UTC) We don't give medical advice. That being said, they recommend a booster tetanus shot every 10 years. Clarityfiend (talk) 22:56, 27 November 2022 (UTC) Okay. 67.215.28.226 (talk) 23:01, 27 November 2022 (UTC) Tdap and any similar vaccine cannot prevent the disease if a person is already infected. Your mother may need tetanus immunoglobulin, not the vaccine. Ruslik_Zero 19:56, 28 November 2022 (UTC)