Wikipedia:Reference desk/Archives/Science/2015 August 5

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

August 5

Can anyone identify this plant?

Found near Whitney Portal, July, 2015. --Trovatore (talk) 02:08, 5 August 2015 (UTC)

I am not good at plant identification, but Flora of the Sierra Nevada alpine zone is a good starting point. Do you recall at which approximate altitude you found the plant? The biomes change dramatically in the Sierras.

If you found it at lower or moderately high altitudes, it might be on List of plants of the Sierra Nevada (U.S.). (It looks like your photo is still in a forested zone).

Nimur (talk) 02:42, 5 August 2015 (UTC)

Roughly 2900 m (9500 ft), if that helps. --Trovatore (talk) 03:07, 5 August 2015 (UTC)

I'm going with cow parsnip - but please recognize that I am not an expert plant identifier. The leaves are not an exactly perfect match, but the general appearance is similar.

I came to this conclusion by pulling out a copy of my not-quite-field-portable guide book, The High Sierra (1972). It has an entire chapter on plants of the high sierra. Nimur (talk) 03:27, 5 August 2015 (UTC)

Ah, thanks much. Could well be. It certainly appears to be one of the Apiaceae in any case, which includes some harmless plants and some very nasty customers. For what it's worth, I tasted it, and I didn't die, though that doesn't prove much as I could hardly have ingested more than microgram quantities. It was not unpleasant, rather like pine needles. But I don't recommend anyone else repeat the experiment. --Trovatore (talk) 03:41, 5 August 2015 (UTC)

Strange. I was in this area four years ago and didn't find this plant.--Jasper Deng (talk) 09:29, 5 August 2015 (UTC)

• Definitely one of the Apiaceae, but without a clear picture of the leaves it's very difficult to pin down. Cow parsnip has quite large leaves, and I can't see any sign of them in the picture. One strong possibility is Angelica, specifically Angelica lineariloba (also known as "poison angelica", hint hint), which is widespread in that area. Looie496 (talk) 13:31, 5 August 2015 (UTC)

I was going to say angelica, but I was thinking of the edible kind - Angelica sylvestris - but I think you've hit the nail on the head. Alansplodge (talk) 17:35, 5 August 2015 (UTC)

Tasting it is certainly not a good idea. Although it's impossible to tell the size of the plant from the photo, it looks a quite like (and is at least related to) a plant that is spreading across many parts of the world, including the US, known as the Giant hogweed whose sap can cause severe burns with long lasting effects. It tends to attract attention as it grows very tall but is best avoided.[1] Richerman (talk) 09:00, 6 August 2015 (UTC)

The "feathery foliage" at the bottom of the photo matches poison angelica but is totally dissimilar to giant hogweed. Alansplodge (talk) 20:58, 7 August 2015 (UTC)

geometric plane

hi, i want to find some geometric detail as described below: i need to find the equation for displacement in centroid of a triangle on displacing each of vertex of triangle independently in z direction. please tell me,where to get reference material. SD — Preceding unsigned comment added by Sameerdubey.sbp (talk • contribs) 05:59, 5 August 2015 (UTC)

If I've understood you correctly, and the triangle is in the XY plane, then the centroid just moves one third of the displacement of the vertex, since the centroid is two-thirds of the way down each median. I expect that someone else can give you a more general formula with references. Dbfirs 07:04, 5 August 2015 (UTC)

...which means that if all three vertices move in Z, then the centroid is displaced by one third of the sum of the three vertex movements - which is more easily stated as "the average of the three vertex movements" - which is kinda what you'd hope it would do! SteveBaker (talk) 14:42, 5 August 2015 (UTC)

• The coordinates of the centroid are the average of the coordinates of the vertices; it follows that the movement of the centroid is the average of the movements of the vertices. (For future reference, you might note that we also have a Mathematics Desk.) —Tamfang (talk) 04:31, 6 August 2015 (UTC)

Is the length of a river well-defined?

I was looking at the list of rivers by length and found that there appears to be no concrete definition of the length of a river. The length of the shorelines of a river is not well-defined because of the coastline paradox. Because river channels are, in general, quite irregular in shape, especially near the sources, I think it is futile to define the length of a river.--Jasper Deng (talk) 09:28, 5 August 2015 (UTC)

The length of a river cannot be measured exactly. This is discussed and explained at List of rivers by length#Definition of length, which you appear to have read.--Shantavira|^{feed me} 12:32, 5 August 2015 (UTC)

The question "is the length of a river (or coastline) defined to infinite precision", the answer (because of the coastline paradox) is "no". However, you can define a river (or coastline) length by constraining your definition to certain tolerances and resolutions, for example by creating an algorithm that maps a series of line segments of defined length onto the course in question; the line segments can then be summed up, and a useful and meaningful length can be created; you just need to know what definition and algorithm you're working with so you know how the result was obtained; the result of a good algorithm should, within any reasonable precision, provide an answer which matches the sort of physical experience people would expect. For example, people who want to know the length of a coastline really probably want to know "If I walked the entire coastline, how far would I walk." If your algorithm produces the same result as the experiment if you actually walked it out, then it's a useful one. A simpler way to think of it is like trying to apply the Koch snowflake to a much simpler polygon. Yes, the Koch snowflake is an interesting idea (in the same way the coastline paradox is), but the Koch snowflake does not mean that we can never give a meaningful answer for the perimeter of a polygon. The same with the river length/coastline paradox issue. --Jayron32 12:44, 5 August 2015 (UTC)

No, the "If you walked the coastline" test doesn't work because it depends entirely on how closely the person followed it, whether they walked into every tiny inlet and out onto every prominence - it has the very same fractal problem that is discussed in the coastline problem.

One solid definition for coastlines is the convex hull - which is consistent, well defined - but highly unsatisfactory for coastlines with large-scale concavities. But for rivers, one might take a similar approach by hammering a stake into the riverbed at the source, tying a rope to it and letting the rope follow the river to the center of the mouth. Pulling the rope as tightly as possible without it disturbing the banks would provide an unambiguous measure of "length" - presuming that you have a solid definition of which tributary is the "true source" and which exit to the ocean is the final end of the river, that the path of the river through tributaries is agreed and that the rope is infinitely thin and flexible.

That's not so easy in practice. Consider the mouth of the River Thames in England - which doesn't have a very clear and obvious transition into "ocean" - is the "end" at the mid-point of the tip of the Isle of Thanet and Foulness Island - or is it more like halfway between Southend and Sheerness? We could make a definition for the mouth such as "when the outflow crosses the convex hull of the surrounding land"...but that's not particularly acceptable when an entire continent has a large concavity such as with Rio Lluta at the border of Peru and Bolivia.

But if you have those definitions, then as a theoretical approach, the "rope stretching" thought experiment gives you the shortest path from source to mouth - which is a reasonable, finite and consistent measure. However, it's doubtful that this approach is practical in reality because you can't stretch the rope in that manner without having to worry about waterfalls, rocks that stick up out of the water and the curvature of the earth - to pick just a few examples - so if you can't do it in practice - you're entirely reliant on the precision of maps - and those suffer from the exact same fractal problem. SteveBaker (talk) 14:39, 5 August 2015 (UTC)

Only if you leave the definition of the walk undefined. Again, you can create a practical definition for "how closely to follow the coastline", and that will give you a real answer which is really useful. The paradox only comes into play where you leave your definition at infinite precision; it's like the Koch snowflake problem again: once you define a polygon as having a finite number of sides, you can find a real perimeter. It only becomes an undefined (or infinite length) perimeter if you demand that the Koch snowflake have an infinite number of sides. Your coastline is only undefined if you demand that one follow the coastline to infinite precision. If you allow the edges under a certain dimension to be rounded off, and allow the walker to "skip" following perterbations under a certain scale, you can define the coastline of a certain length. The coastline paradox is a classic example of where the usefulness of a mathematical model breaks down in the face of a real application. If I launch a boat in Oxford, and I want to know how far the boat will have traveled by the time it gets to London, I can do that so long as I define the trip to a reasonable (instead of infinite) level of precision. --Jayron32 15:09, 5 August 2015 (UTC)

The ambiguity of a coastline length and the ambiguity of a river length seem to be different issues. The length of a coastline is difficult (or impossible) to measure, because you can always zoom in closer and find more convoluted indentations to increase the overall lenght. A river doesn't have that problem, at least not if you measure its length along the median line. Infinitely convoluting the banks won't increase the overall length of the median line. The ambiguity of a river's length seems to be due to arguments about which tributaries to include, how to define the start when the river may be e.g. gradually emerging from a wetland, and actual physical changes in length or course of the river (not all or any of which need apply in all cases). Iapetus (talk) 15:23, 5 August 2015 (UTC)

Infinite series can converge. In other words, even if the coastline or the river (or the abstract mathematical path) has infinite complexity, it may have a finite and well-defined length. This is possible if certain specific mathematical properties of the path are conserved. It so happens that these properties match our intuitive expectations of a real coastline, or a real river path! In other words, there is no "paradox" in the coastline length. If you apply a mathematical model incorrectly, you will compute a wrong (or infinite, or undefined) result... but that ought to clue you in to the fact that you're not using the correct model!

Real coastlines and rivers have complex shapes: but even if we trace the paths ad infinitum to describe their smallest, most minute details, they still are not true fractals in the absolute mathematical sense. Fractals have specific mathematical properties that are described by recurrence relations. The perimeter or path-length for some fractals does indeed diverge (i.e. the length is not well-defined). However, it is a terrible and incorrect myth to assume that any and all complex shapes are fractal! There are many cases of non-fractals in analytic geometry, where we can study infinite numbers of infinitesimal features - like a river squiggling its way through a path, or a coastline curving and squiggling along - and the integral of the length over all these tiny path excursions need not be divergent. We can work with infinities and limits and infinite quantities of very tiny curves - modern mathematics provides us with lots of tools for that - but we must do so carefully in order to avoid incorrect conclusions.

I come from a background of digital signal processing. When I look at measurement of a pure mathematical curve, I know its exact equation; but when I consider the curve that approximates a shape in nature - like a waveform or a path of a river on a map - I am always thinking about a few ideas: in particular, I think about the sampling theorem, and I think about spatial filtering. The river's path is not an exact analytical curve defined by an equation that we already know. So, we are not actually measuring the length of a river: we are, in fact, computing the length of a river, whose path sampled at the spatial resolution of our map, and filtered by our smoothing functions (in geographic information systems, we use spline equations fitted to the data, which in essence is the application of a type of low pass filter). We can know whether these preprocessing steps introduce errors: we can resample at a lower (coarser) granularity. If the result is dramatically different, we have insufficient sampling. By the time we have appropriately sampled the problem, we compute approximately the same result - even if we cut the sampling rate by 2x or 10x. In other words, the length of the river (or coastline, or path) has converged. This is a numerical computation trick that works across the board in many domains where our model of the real world is discretized. At its core, this is the same mathematics as Von Neumann stability analysis - we are checking if we have created a numerical error by choosing an inappropriate value for the differential path element.

All these ideas - most particularly, the difference between "defining" and "measuring" and "computing" - are swirling around... and these are actually really important topics to explore. Mathematicians, geographers, and computer programmers mean something different when they say words like: define, "well-defined," "measure," "length," ... so you'll get a different answer depending on who you ask.

Nimur (talk) 15:36, 5 August 2015 (UTC)

@Wardog: I'm not so sure, because then how can I be guaranteed that the median line itself is rectifiable? To me this line would be defined by the midpoint between the banks, and I don't think this would be any smoother than the shorelines.--Jasper Deng (talk) 17:51, 5 August 2015 (UTC)

Yeah - I agree. A small pebble on one side of the river still perturbs the mid-line by half the width of the pebble - and river banks are just as fractal as coastlines (the fractal dimension is likely to be different - but the resulting length is still infinite). The problem of defining how small of a feature is counted still applies - and without counting, the length of the median is still infinite. It's precisely the same as the coastline problem...except that, with a coastline, the "convex hull" (which is well defined) seems an inadequate measure - where the 'tightly stretched rope' approach that I described doesn't seem so horrible in the case of a river.

What fails in every case (for both rivers and coastlines) is that people insist on telling you the lengths of those things WITHOUT adequate stipulation of the definition for what is being measured. That's weird because nobody ever tries to tell you the circumference of a cloud - and that's the same exact problem. SteveBaker (talk) 18:10, 5 August 2015 (UTC)

@SteveBaker: That's not quite right. The median line gets effectively smoothed by the width of the river. If you have a river with some minimum width w, then the radius of curvature of the median line won't be smaller than something like w/2. You have to put the median line equidistant from the nearest point on each shore. --Amble (talk) 19:21, 5 August 2015 (UTC)

I want to reiterate: the paths that follow the edges coastlines and rivers are not fractals. Even our article, coastline paradox, explains that they are "fractal-like." These paths have properties that are fractal-like; but fractals are non-differentiable. That property is a very important one: it's the key to the whole "fractals have no well-defined length" paradox. Does that specific mathematical property - the non-differentiability - apply to the edge of a river bank, or to the path between the sea and the land? If that non-differentiability property does not apply, then the path is not a proper fractal. The ambiguity in defining the path length, therefore, is not because the length is infinite. The ambiguity ultimately stems from some other error in the construction of the problem. At best, these natural shapes are convenient analogies for the mathematical abstraction that is a fractal curve.

Let's especially distinguish between a path whose length is infinite, and a path whose length is ambiguous; and distinguish both of these cases from a natural shape whose path is not easy to describe with an analytical equation. These are all completely different concepts.

Nimur (talk) 19:39, 5 August 2015 (UTC)

An arc does not even need to be differentiable to have a well-defined length - the definition at arc length specifically does not require it. But coastlines (including river banks) do fail to even be continuous, let alone differentiable. If you look at the boundary between the water and the land, it is not a single line. Waves and water levels change that too, and then when a tributary enters, where do you demarcate where the tributary ends and the mainstem begins? Hence the median line cannot be considered to be smooth either. The concept of a median line also becomes ill-defined for braided streams.--Jasper Deng (talk) 22:09, 5 August 2015 (UTC)

@Amble: Also, that is not of any use because the minimum width of a river at its source is usually a very small length. And that still leaves ambiguity of where to place the median line when the river is substantially wider than w.--Jasper Deng (talk) 22:15, 5 August 2015 (UTC)

@Jasper Deng: You can always inject some degree of uncertainty, but the specific claim I was responding to -- that the median line is some untameable fractal beast with infinite length -- is simply incorrect. The indentation level of my posting indicates threading per WP:INDENT, but since this rule isn't much followed these days, I suppose I need to train myself to explicitly include a ping or username every time. --Amble (talk) 22:32, 5 August 2015 (UTC)

Given two agreed upon points in the water, there is a well-defined minimum possible path length between them subject to the restriction that the path never leaves the water (or the water as it exists at some fixed instant in time anyway). For me, that seems like a reasonable definition for the length of a river provided one can agree on start and end points (which can be a rather thorny issue). This speaks to Steve's rope stretching analogy, and essentially gives the minimum distance a traveler must cover if moving along the river from start to end, which seems like a useful notion. Of course, in practice, river lengths are really always measured by creating a map with some finite level of detail and then measuring the course of the river on the map, in which case the answer will depend on the precision of the map. A cartographer might also try to create a fancy mid-river path rather than the shortest within river path, but that's mostly an issue of preference. Dragons flight (talk) 22:28, 5 August 2015 (UTC)

5 Alpha Reductase

There are many 5 Alpha Reductase inhibitors in natural organic forms. Do they really inhibit 5AR enough to decrease Di Hydro Testosterone as a result ?? — Preceding unsigned comment added by 103.15.60.58 (talk) 12:11, 5 August 2015 (UTC)

5-alpha-reductase inhibitor discusses testosterone inhibition and use in male and female pattern hair loss. Though our article mentions a large array of plant-derived inhibitors, it currently calls their effectiveness "unknown", and frankly, I don't want to take the time to run down every single one in PubMed right now to see if any further data is in about them. Wnt (talk) 16:23, 5 August 2015 (UTC)

Kidney transplant life expectancy

Hello, What was the life expectancy for a kidney transplant patient in 1975? Thanks, Chris. — Preceding unsigned comment added by 31.52.165.152 (talk) 18:06, 5 August 2015 (UTC)

According to this study the life expectancy in 1989 in Scotland was 17.19 years vs 5.84 for those on dialysis. Ruslik_Zero 19:26, 5 August 2015 (UTC)

Nylander's test

Do we know after whom Nylander's test is named? DuncanHill (talk) 22:28, 5 August 2015 (UTC)

Based on [2], I believe the answer is Swedish chemist Emil Nylander (1835–1907). Dragons flight (talk) 23:32, 5 August 2015 (UTC)

I am finding mostly Almen-Nylander test, and one Böttger-Almen-Nylander test. sv:August Almén maybe?—eric 01:42, 6 August 2015 (UTC)

de:Claus Wilhelm Gabriel Nylander? —eric 01:47, 6 August 2015 (UTC)

de:Nylanders Reagenz agrees that his work was the final stage in an evolution of tests based on work of Böttger and others. And that he published the work under the name "Emil Nylander". The birth/death years in the DEWP article Eric linked match what Dragons flight mentions, supporting that they are the same person (I just created a REDIRECT on DEWP for it). DMacks (talk) 09:13, 6 August 2015 (UTC)

Rudolf Christian Böttger.—eric 02:03, 6 August 2015 (UTC)