Wikipedia:Reference desk/Archives/Science/2007 December 29

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.

December 29

Amount of H2O throughout Earth's history?

Are there estimates of how much water has been created or lost throughout earth's history? I'm not talking about flowing water. I'm asking about the water on earth in whatever phase it is. 128.163.7.237 (talk) 00:36, 29 December 2007 (UTC)

To be honest it's a bit difficult to know where to start, which might be why you haven't received any replies. Most of this stuff is easy to find out using Google, for example.

The water on Earth came here via comets over the hundreds of millions of years after the planet formed around 4.6 billion years ago. The amount of water on Earth is not constant as some of the hydrogen and oxygen will be bound into other things or get lost into space, but we also get more of these elements *from* space so that evens out the equation pretty much.

No matter how you use water on Earth it remains in the system, because human made things last extremely short compared to the time scales of geology. So eventually water you mix into concrete will escape and make it back into the ocean, for example. This process can take a long time, but all water on Earth is part of a huge cycle.

Water is perhaps the most important molecule for life on Earth, when we consider that absolutely all life as we know it consists mostly of water (our bodies are a prime example). Since we are a part of the ecosystem on Earth we do not "remove" water from the cycle. Every time you breathe or move, water vapor is lost from your body into the atmosphere. Urine is a waste product from metabolism and as such does not take away nor contribute, since all the ingredients are natural and come from outside your body in the first place.

So I guess the answer to your first question is, "no, in the beginning of the planet's history there was no water, but we have evidence of life in water that is about 3,6 billion years old, so we know that water has been around for a very long time".

If the temperature on Earth changes dramatically, the water will condense into thick clouds but it will likely remain on the planet. If for some reason our atmosphere should thin out, the water will boil off the surface and evaporate into space. --n1yaNt 01:39, 29 December 2007 (UTC)

If Hydrogen gets lost from the atmosphere then that might effectively reduce the amount of water on earth. Polypipe Wrangler (talk) 04:46, 29 December 2007 (UTC)

Actually there are indications from the Hadean era that oceans existed shortly after the planet formed. This is expected because the rocks forming the planet would have a lot of hydrogen and oxygen. No need to wait for cometary water to get oceans. I don't know of studies of how much hydrogen our atmosphere is leaking, but there are people trying to figure out how much water we are now losing and gaining. We'd have gained some from various incoming rocks, even the tiny ones we are always dusted with. Our water cycle is relatively closed, but some leakage is suspected. -- SEWilco (talk) 06:28, 29 December 2007 (UTC)

It also depends on what you mean by "created" and "lost". If you run electricity through water, you'll split it into hydrogen and oxygen gas. If you then burn the hydrogen, you'll get the same amount of water back as you actually split. Was that water "created" from the hydrogen, even though there was no net change in the total amount of water on earth? Same thing happens with fossil fuels - the hydrogen in the hydrocarbons combines with oxygen in the air to form water. Is that water "created", or are you simply restoring the water that was taken up by the prehistoric plants and animals that were transformed into the oil? During the course of a normal day of metabolism your body (and that of plants and other animals) breaks apart and puts back together millions and billions of water molecules. If you get really picky, water molecules just sitting in a glass aren't static, either. Even ultrapure water is constantly breaking apart (temporarily) into hydrogen ions and hydroxide ions. The hydrogen ions form hydronium ions, and the hydrogen atoms get switched around and swapped rapidly, before a different hydrogen recombines with the hydroxide ion. If you were to label an individual oxygen and the two hydrogen atoms attached to it, you'd see that within seconds the two hydrogens would go drifting off to different oxygen atoms, and the labeled oxygen would get new hydrogen partners. Probably not what you meant, but something to think about when you think about creating and destroying water. -- 18:30, 29 December 2007 (UTC) —Preceding unsigned comment added by 128.104.112.236 (talk)

Thanks for the responses, everyone. Regarding the question above: In the case of water formed from from separate H and O atoms, I would think there's net water created, even though the net amounts of each H and O molecules stay the same. In the example of pure water's acidic behavior, it's true water is both "created" and "lost", but in this case the equilibrium ensures that there's no net loss or gain (unless the system is disturbed).

I'm interested in particular the reactions in the early geological history that created the abundance of water, as well as possible reactions that could or did cause net loss of water. Since water is the essence for life, the processes that make it are interesting to know.128.163.7.250 (talk) 13:09, 30 December 2007 (UTC)

Backcrossing

MyD88-/- mice were generated as described and backcrossed for 9 generations on an H-2d (BALB/c) background. - what's the backcrossing necessary for? --Seans Potato Business 20:15, 12 December 2007 (UTC)

When generating knockout mice, it's generally necessary to mix two different strains of mice. Backcrossing is used to reduce the genetic components of one of the founding strains. — [[Scientizzle 21:16, 12 December 2007 (UTC)

Here is an ugly graphical representation. After several generations of backcrossing, one can assume that the only non-BALB/c DNA in these mice is from the regions flanking the knocked-out gene — Scientizzle 21:21, 12 December 2007 (UTC)

That's great; thanks. So you only need to perform backcrossing when you have the knockout in a different genetic background than the one you want? Would it not be simpler to have performed the knockout in the genetic background in which you wanted it in the first place? Are some backgrounds in some way easier to use for producing knockouts? --Seans Potato Business 22:07, 13 December 2007 (UTC)

The formation of a genetic knockout mouse is moderately complex. A genetically modified embryonic stem cell of strain 1 is placed in the developing blastocyst of strain 2. In order to determine the efficacy of the implantation (and to plan subsequent breeding), it's useful to produce a chimera of strains with different coat colors. There's more information at Knockout mouse. — Scientizzle 22:22, 13 December 2007 (UTC)

But the diagram you link to uses three different strains. One cell type endures the knockout and is put into a blastocyst of another type. They then try to get the transgene in a C57BL/6 background by "backcrossing". If they (the scientists involved) had used C57BL/6 cells in the first place, they could put those cells in the same blastocyst as before, breed with a C57BL/6 mouse, and just like that, all the black offspring are heterozygous for the knockout and fully C57BL/6; no backcrossing necessary. --Seans Potato Business 13:30, 15 December 2007 (UTC)

As far as I can tell that is what this diagram shows. The founder is a chimera of C57BL/6 (abbreviated as B6) and 129/Sv cells and it is backcrossed to C57BL/6 mice. That is just two strains. However, you don't put B6 cells into a B6 blastocyst, because then its is much more difficult to identify which offspring are chimeric. Typically need your null allele to be incorporated into the DNA from a strain with a different coat colour, since this is the way you can identify your founders (129 mice are white, and B6 mice are black - chimeric mice will be a patchwork of white and black). So at N1, the first backcross, you will select mice that have one allele that is B6 and the other has to be the null allele, which is on a 129 background. Thus, at that point, you will have a 50/50 mixed genetic background. Rockpocket 19:35, 21 December 2007 (UTC)

Having a coat color chimera allows for a greater simplicity in creating a knockout strain--the chimeric progeny of a successful incorporation will be obvious, they're adorable little multicolored animals with the coat colors of both the recipient cell and the donor cell (the cell w/ the altered DNA). Those that have the color type of the donor in higher ratio, and especially if that coat color is found near the reproductive organs, are more likely to be heterozygous for the target mutation in their germ cells, which is the key. The progeny of those animals become your F1 generation if germline transmission is received. Because raising animals is expensive, and genotyping a new mutation can be difficult/expensive/time consuming, these shortcuts make it simpler to produce a strain.

One note--the mice I've been involved with ususally only use two strains, 129/SvEv cells injected into C57BL/6J, then further backcrossed onto C57BL/6J, for example. Why one would use three strains, I don't know...but using only one strain would cause large difficulties...it would be effectively impossible to determine the efficacy of any blastocyst implantation; one would be left with attempting to test every possible F1 pup for the genotype of interest. — Scientizzle 20:35, 21 December 2007 (UTC)

But aren't you assuming that the coat colour is linked to the chromosome holding the knockout? In reality, they will probably separate, being on different chromosomes, leaving a mouse with grey colour but homozygous positive (no knockout), indistinguishable from a mouse that has the same colour and is heterozygous.

Scheme:

Chimeric + White =

Grey hetero

Grey hetero + White =

Light grey hetero and light grey homo +ve

--Seans Potato Business 23:07, 21 December 2007 (UTC)

No.

1. DNA-altered ES cell from gray-coat-colored strain added to developing blastocyst of a black-coat-colored strain.
2. Resulting offspring are chimeric: stripes & patches of black and gray. The level of gray coloration is suggestive of the extent of the expansion of the altered cell into the various tissues, including the germ tissues.
3. Those mice are crossed with black mice. Those offspring will only be heterozygous for the mutation if their chimeric parent's germline cells were derived from the altered (gray) stem cell. Coat color may be, at this point, genetically independent of the mutation, as you suggested. (But! pups of variable color--half completely black and half of the litter completely gray--are suggestive of gray-type DNA in the germ cells, so that's usually a good sign.)
4. These putative F1 mice are then genotyped...and the rest is history.

The key to the whole process is that the genetically altered stem cell must end up as a progenitor to germline cells in order for the whole scheme to work. Mixing coat colors is a way to measure that efficacy. I'm not saying this is the best way to do it, but it is why it's done. If one only uses C57 cells in C57 embryos, one cannot easily know if any step (incorporation of ES cell, expansion of that cell into relevant tissue, inheritance of said cell's coat color phenotype) along the way, until genotyping the F1, has worked. — Scientizzle 23:59, 21 December 2007 (UTC)

Okay, I understand it. I'll try to make a nice SVG diagram for the article. Here's a method that I think would be better though; it requires approximately eight genotypings but no backcrossing. Is it flawed?

1. Transduce C57 cells and insert in 129 blastocyst
2. Cross chimera with C57 mice
3. Black offspring are heterozygous
4. Cross black offspring
5. Genotype approximately four males and four females to find one double negative of each type.
6. Breed as much as you like. They're all fully C57 background and so no backcrossing required. --Seans Potato Business 09:52, 22 December 2007 (UTC)

Sorry Sean, your scheme doesn't look too good. #3 is wrong because, unless the mutation of interest is on the same chromosome as the coat color gene(s), one cannot assume that the black offspring are the hets. In fact, if unlinked, black offspring have same odds of heterozygosity as non-black.

The backcrossing is necessary because any het F1s would, by definition, have ~50% 129 strain DNA. (The chimera parent has a mostly C57 body, but must have 129 sperm or eggs in order to contain the mutation.) Each F1 het would have a different profile of C57 and 129 DNA, too, so crossing them could make offspring that are between 0 and 100% genetically 129. Continued backcrossing against C57 +/+ animals reduces, each generation, the amount of (mutated-chromosome-unlinked) 129 DNA by 50%, until only the region of the altered chromosome is non-C57. This is important because different inbred mouse strains can have wide variability in many important features, including behavioral measures, drug metabolism, cancer development, etc. — Scientizzle 20:53, 22 December 2007 (UTC)

Okay, I messed up on the first scheme, but I'm not through yet!

=== Revolutionary Nobel-prize-winning method ===

1. Transduce C57 cells and insert in 129 blastocyst
2. Cross chimera with mouse of C57 background
3. 50% of black offspring are heterozygous - genotype approximately two males and two females to find one heterozygous of each gender
4. Cross heterozygous black offspring
5. 25% of offspring are homozygous negative - genotype approximately four males and four females to find one homozygous knockout of each gender
6. Breed as much as you like. They're all fully C57 background and so no backcrossing required

Stats to achieve two homo.neg. in required background: Total crossings = 2, total genotypings: 12

=== Standard Leading-brand method ===

1. Transduce 129 cells and insert in C57 blastocyst
2. Cross chimera with mouse of C57 background
3. 50% of not-black offspring are heterozygous - genotype approximately two mice to find a heterozygous negative mouse
4. Cross heterozygous negative with mouse of C57 background
5. Offspring are all about 75% C57 and 50% are heterozygous - genotype approximately two mice to find a heterozygous negative mouse
6. Repeat steps 4 and 5 10 more times to achieve mice with 99.98% C57 genome and genotype approximately two male and two female mice to find one heterozygous mouse of each gender
7. Cross these two mice to achieve 25% homozygous knockout mice with 99.99% C57 genome - genotype approximately four males and four females to find one homozygous knockout of each gender

Stats to achieve two homo.neg. in required background: Total crossings = 13, total genotypings: 32

Am I right in my understanding of the backcrossing procedure and/or my assertion that reversing the roles of the C57 and 129 mice is a sound suggestion? --Seans Potato Business 16:10, 25 December 2007 (UTC)

Not really. Your idea of reversing the roles of C27 and 129 is flawed because of the dominance relationship of alleles of the tyrosinase gene. The Tyr⁺ allele is dominant to the Tyr^- allele. 129 mice are Tyr^-/- which makes them albino. C57 mice are Tyr^+/+ which makes them melanic. At step 2, all the offspring from the cross will be melanic (either Tyr^+/+ or Tyr^+/-) Therefore there is no way of knowing, by looking at the offspring from step 2, whether the transgene went germline in the chimera. As Scientizzle notes above, the key to the whole process is that the genetically altered stem cell must end up as a progenitor to germline cells in order for the whole scheme to work, your suggestion does not help us determine whether that step has occurred. Rockpocket 05:34, 29 December 2007 (UTC)

Thinking about this further, I fear I have given you misleading information. While my points are correct in reference to the original 129/SvJ strain, it turns out that typical 129 available mice today are actually from a congenic substrain called 129/Sv-Sl/+,+^p,+^c. The coat colour in this mouse is from a mutation at the Steel locus, giving the mouse a "steely" white/grey colour. Importantly, this is dominant to the C57 Steel allele, so if one were to use this substrain then my criticism is no longer valid. However, there is another technical problem, and that is that pretty much all the embryonic stem cells lines widely available (in which you need to make your knock-in/out) are from the 129 strain. Again, there are a number of technical and historical reasons for this. If you could get adequate C57 ES lines, then I think your strategy would work. However, as with most revolutionary Nobel-prize-winning methods, you are not the first to think it up - the Gene Targeting Facility at the University of Virgina already offer this service using their own source of C57 derived ES cell lines. [1] Rockpocket 09:33, 29 December 2007 (UTC)

Okay, thanks for the insight. I've made a schematic for the backcrossing article. Comments? --Seans Potato Business 14:11, 1 January 2008 (UTC)

I don't see anything at that link, I'm afraid. Rockpocket 03:06, 2 January 2008 (UTC)

Apparently you need to click on another link to get to the actual file. It confused me for a while, too, since the page you get at first looks just like a typical parked domain.

Anyway, the picture seems basically correct (and rather funny, too) to my untrained eye. It's somewhat confusing though — I had to look at it several times to figure out which mouse was the offspring of which mating. Perhaps it would be better to use more whitespace, and to draw an explicit family tree with lines connecting each mated pair and pointing to each of their offspring. —Ilmari Karonen (talk) 17:24, 2 January 2008 (UTC)

Here is Mk II. Is it any better family-tree-wise? Do you still think it needs white space, and if so, between which elements exactly? Your input is valuable and appreciated. ----Seans Potato Business 00:38, 3 January 2008 (UTC)

That's weird...we answered this question on Dec 12th. What's it doing back here again? The OP's signature is even from Dec 12th. SteveBaker (talk) 21:04, 21 December 2007 (UTC)

Sean brought it back for round 2... — Scientizzle 21:09, 21 December 2007 (UTC)

RockPocket brought it back for Round 3 now? Are we attempting to make this the thread that never dies? Should we give it its own subpage? ;-) Nil Einne (talk) 08:53, 29 December 2007 (UTC)

Not me. This is Sean's baby. Rockpocket 08:57, 29 December 2007 (UTC)

Ooops sorry then, I thought it was you who brought it back with your response but I guess Sean brought it back because he'd received no response Nil Einne (talk) 09:08, 29 December 2007 (UTC)

Watermelon snow SOLVED

Resolved

I saw a post in the recent archives about Watermelon snow and, fondly recalling a summer spent with the USGS, went to look at the page. When I first encountered Watermelon snow, (In I think Lassen Volcanic National Park?) the guy in charge told us that we could smell it but that we should never eat it because it was toxic to humans. I see on our page that it is just a type of green algae. I think if it *is* toxic, that should be mentioned on the page. But interestingly there is no mention on the page of either toxicity, or conversely, what it tastes like. Anyone know the truth?

I should add that reports on the internet vary and seem anecdotal. Saudade7 14:41, 29 December 2007 (UTC)

See WP:BEANS. Most stuff is toxic. In fact everything is toxic if you ingest enough of it. That doesn't mean that everything in Wikipedia has to be labelled toxic. If the guy in charge of it told you it was toxic, that's good enough for me. However, if you really think some dork might eat it, by all means add a warning in the article. I guess it's then up to whoever removes the comment to show that it isn't toxic.--Shantavira|^{feed me} 17:02, 29 December 2007 (UTC)

So someone who has devoted his life to Parks and Recreation tells you it is toxic and you don't believe him? Huh. Just take the benefit of the doubt Saudade7. n1yaNt 17:59, 29 December 2007 (UTC)

Ha ha! You *do* make me sound stupid! But for my part, I guess I have always been one of those "question authority" types who wants to see some kind of O-Chem proof before I believe someone! And also the "guy in charge" worked for Florida International University as a geologist. He didn't work for the Parks department, so I think he could be wrong. Also, while it may be the case that most stuff is toxic, esp. "if you ingest enough of it" (e.g. water), I think something like Watermelon snow that smells fruity and delicious, and has a delicious sounding name, should probably be mentioned as toxic if you don't have to ingest significant amounts to experience deleterious effects! Also, I cannot use my heresay as a <ref>! So I just wanted to know if someone actually knew the verifiable truth. Thanks! Saudade7 19:33, 29 December 2007 (UTC)

Googling about comes up with a lot of "I don't know if it is toxic for sure but it is probably not good for you at all and may very well be toxic." Cecil at the Straight Dope has a posting about it; he's inconclusive. There have been, of course, [http://www.pesticideinfo.org/List_AquireAll.jsp?Species=980 toxicity studies] on the algae in question (Chlamydomonas nivalis), though I'm not acquainted with the metrics they use there and can't derive much from it. As for whether a park ranger would know—who knows. People repeat all sorts of things they have heard, and obviously most people aren't interested in being the person to try it to debunk it (or prove it right). If they haven't conducted toxicity studies or have read the relevant literature then their opinion is easily up for some doubt, sure. --24.147.86.187 (talk) 19:40, 29 December 2007 (UTC)

Yeah 24.147.86.187, I figured that if the Straight Dope entry was inconclusive, that was a bad sign!! Thanks! Saudade7 19:43, 29 December 2007 (UTC)

There's a paper on the subject from Wilderness and Environmental Medicine Vol 8 issue 2 here. William Avery (talk) 19:56, 29 December 2007 (UTC)

Well, that answered that! (Not toxic - no greater frequency of diarrhea than placebo.)

Thanks William Avery !!! Saudade7 20:48, 29 December 2007 (UTC)

I bet they had fun writing that grant application! The way I'd have handled it is, get the research team together, do tequila shots, write the grant, and throw that one away. Then write the real application a few days later, stone sober, with all the jokes out of the system.

The last paragraph of the paper suggests that the next application is already in the works.... Trovatore (talk) 10:16, 30 December 2007 (UTC)

Well that was in 1997 and presuming there's nothing new I would guess their next application failed. BTW, they only studied diarrhea and as they said, 500g each of snow in one location so I don't think it definitely answers anything. Arguably they would have observed any significant acute effects due to 500g of snow but they definitely they wouldn't have observed chronic effects. To use another example humans are apparently somewhat resistant to the acute effects of Aflatoxin and I don't know how easy it is to get diarrhea from consumption of aflatoxin contaminated food but I suspect not that easy. Yet it is definitely a toxin given the chronic carcinogenic effect and many people particularly in the developing world IIRC are affected somewhat. (This link [2] confirms my suspicions) Of course unless you are intending to regularly consume red snow you probably don't have to worry about chronic effects but it's important to consider what the study did show and what it didn't show and it definitely didn't show that red snow isn't toxic. P.S. They also only used 7 volunteers, not exactly a large sample size. P.P.S. The issue here of course is that people don't regularly consume large amounts of red snow so no one particularly cares if red snow can really be considered toxic or not. The only thing which is of marginal interest is will just trying a bit of red snow cause any significant problems and the answer appears to be no. Nil Einne (talk) 11:44, 30 December 2007 (UTC)

time travel: looking for scientific-sounding mechanism for a SPECIFIC (new?) sci-fi idea

I had an idea of a science fiction technology regarding time travel, set sometime in the next 100 years, involving time travel. What I'm looking for is a science-sounding explanation of how it works that isn't just treknobabble. Obviously it can be fake, but if it sounded at least possible in some universe, that would be ideal. As an example, the Alderson drive used by Niven and Pournelle is based on postulating a previously unknown force that appears under certain circumstances and then explains how that force would work; it's not real, but it's at least plausible if you accepted some of the initial conditions (those conditions just happen to not be true in our universe).

Anyway, here's the idea: there would be a centralized facility that would allow individuals to rewind time. That is, someone operating the machine could make time go backwards (but not forwards), but only as far back as the creation of the time machine itself. Additionally nothing other than the machine would be in any way "outside of time"—if I rewound time 10 minutes, I would become 10 minutes younger and be transposed to the place I was 10 minutes before. The only catch here, and the only way it would be useful to anyone, is that the machine itself would be able to preserve the brain state of the person who rewound the machine—thus I'd be 10 minutes younger but with a brain state that was 10 minutes older, so I could remember that I had rewound things.

Has this been thought of before? (I didn't see anything like it on the time travel page, but I am not an exceptionally creative individual so it seems unlikely to me that I'm the first person to come up with such a thing.) It seems like a way to do time travel without worrying about paradoxes, as there are no separate timelines except in terms of the mental state of the person operating the machine at that exact moment (they'd have memories of things that hadn't happened yet), but that doesn't seem to me to be a big deal. Now the mental state bit could be explained away as them having the technology to preserve mental states in the first place, and one could probably add a disclaimer that rewinding back TOO far could be dangerous as you'd be trying to transplant ideas into a brain which might have been developing a bit different over time, but anyway that seems relatively straightforward compared to the physics problem of time rewinding. (I'd probably have it so that anyone who wanted to use the machine needed some sort of complicated neural implant, which would be in charge of maintaining their brain state. It would also make it so that you couldn't go back any further than when you got the implant added, again making sure no paradoxes would come up.)

OK—so how does it work? Wild speculation is encouraged and desired. Criticism is wanted. This is just for fun. Let me know what you think. --Panoptik (talk) 19:31, 29 December 2007 (UTC)

The time machine is the easy part. There have already been multiple ideas for how such a machine can be created, for example, wormholes. The machine only needs to send information back in time to itself, then whatever changes the brain state can take over. Exactly how complete would the change in brain state have to be? If you merely had the memories of the next day implanted in your mind while you were sleeping, it would be far simpler than modifying the whole brain, but make it seem to the person that they woke up in the past immediately after using the machine.

I don't see what you're saying about having no separate timelines. There's the timeline when your brain and body were the same age, ending with your use of the machine, and there's a timeline that begins when your brain instantly ages a certain amount of time. The first timeline is short (after the split), but it's just as distinct as any.

The disclaimer you mentioned would have little real importance, as you could have two people taking turns going back one day and giving each other the message, and that's only necessary if you manage to find a reason you can't just send text back.

The neural implant would be unnecessary as a plot device, as the laws about time travel would probably make it extraordinarily difficult to use. That could even be the reason people are going to do the brain state change instead of sending themselves an email; if your going to jump through all the hoops to use time-travel, why stop at an email?— Daniel 20:33, 29 December 2007 (UTC)

Not sure I follow—I don't think you're distinguishing the specific type of time travel I'm proposing from more generic schemes. The limitations are purposeful because they avoid the paradoxes and plot shenanigans of the other approaches. --Panoptik (talk) 01:18, 31 December 2007 (UTC)

• Paradoxes are still possible in this regime. For example, you could go back in time and kill the inventor of the time machine, but then you'll never have gone back in time so s/he won't be killed, etc. The possibility of paradox is fundamental; even if you could only transfer a single bit of information back in time, you could get a paradox (for example, using a quantum suicide sort of contraption). --Sean 21:31, 29 December 2007 (UTC)

No, you can't go back in time to kill the inventor before they've invented it. As I said, you can only go back in time up until the point of the machine's own construction. So that's not an issue. --Panoptik (talk) 01:18, 31 December 2007 (UTC)

The way to avoid paradoxes is to assume that the Many worlds hypothesis is true. There is no scientific evidence that it's not true - and many scientists (myself included) feel it's the simplest explanation for things like Schrodingers cat. So - you can honestly assert that there are an infinite number of parallel universes - and every teeny-tiny quantum event splits the universe into two copies.

In this view of things, the act of transporting you back in time causes a 'fork' in the universe in which you existed in the past - this is not the fork in which you climbed into the time-machine though. Events from that point onwards proceed normally - so if you kill your own grandfather, he's dead in this fork of the universe - but you're still OK because you were born in the other fork where granddad is still just fine. You can kill off the inventor of the time machine in this universe - and again, no effect in your own universe - so the time machine still exists to send you back in time.

You COULD also go one step further and say that the time machine doesn't literally send you back in time - it merely splits the universe such that there exists a branch in which your body spontaneously appeared from nowhere. Since there are an infinite number of universes - even exceedingly improbable events DO happen in some of those universes. This too is a true statement in known science (if you accept many-worlds). We know that you cannot pin down the position of a sub-atomic particle exactly and it is therefore possible (albeit exceedingly improbable) for a human being exactly like you to materialise out of nowhere with all of your thoughts and memories intact. Since even ridiculously improbable events like this have a non-zero probability of occurring, there must be a universe somewhere in the muultiverse in which this very thing happened at the exact moment at which you tried to send yourself in the time machine. So in a sense, this kind of time travel happens all the time. It must (if we believe the many-worlds hypothesis) have just happened to you. Somewhere in the multiverse, you are standing next to this big-assed dinosaur wondering why you aren't sitting at home reading Wikipedia anymore.

In many-worlds, you also appear in the future - and in the present time but when humans evolved with tentacles instead of fingers - and in present time just the same as you were expecting but where I didn't deliberately misspell the word "antidisestablishmentarianism", and there will be cases where you appear in deep space and immediately die a painful death. Any and all imaginable situations are happening all the time in parallel universes.

The awkward part about this from the point of view of a time-travel book is that no volition was required in order to make this happen - no time machine was needed and (in a sense) no time travel actually occurred. All you really need to invent is some means by which the "you" that is here and now becomes the "you" that just appeared on the "antidisestablishmentarianism" world 10 years ago. It hurts my head to think about that because the me that just did exactly that is right now wondering how it happened!

But if quantum theory and many-worlds is true - then "time travellers" materialise all the time in all sorts of situations. It's just amazingly improbable that any of them happen right here and now in our present universe. SteveBaker (talk) 17:05, 30 December 2007 (UTC)

OK, that's all well and good, but it doesn't seem to apply to what I've proposed at all. I'm not looking for a generic time machine scheme, I've already got the mechanics of it worked out pretty good, I think. I don't think many worlds is at all what I'm interested in—I don't want many worlds, forking timelines, etc. I want a basically coherent, single timeline that can be (with great effort) ran backwards in a very literal sense. If I walked in from a coffee shop and hit the button on the time machine to "rewind" the timeline by 10 minutes, I should find myself back in the coffee shop. --Panoptik (talk) 01:18, 31 December 2007 (UTC)

Then I think you have to give up on the idea of a reasonable explanation and revert to the usual SciFi trick of using technobabble to cover up the fact that you essentially introduced magic into your story! SteveBaker (talk) 17:09, 31 December 2007 (UTC)

I can't see a plausible way that the time machine would work as described - sending everything in the world except your mental state back 10 minutes. And I agree with the other guys that you haven't really got arround the multiple time lines problems. One way to duplicate the effect you are looking for might be to send a signal back in time (rather than rewinding time) The signal would be memories of the events that have happened in the past 10 minutes. The effect would be that the user of 10 minutes ago would be hit by a sudden premonition of the next 10 minutes, and would be equivalent to the one you are looking for, but the mechanism would be different.

Incidently that sort of thing was described by Terry Pratchett in Thief of Time where discworld yetis had a 'save game' power - when they feel they are heading into danger, they can save their life up to a point, and then go on in the knowledge that if they die they will be reincarnated (with memories of the future intact) at the point in their timeline where they last pushed the save button. robcraine

What can be classified as a reducing sugar?

I read the article, and I understood most of it except one part: it says that lactose is a reducing sugar yet on the last paragraph it states that a reducing sugar is not connected with another sugar. Lactose is a disaccharide, so there is two sugars linked together. How can it be a reducing sugar?

If I could get an answer that would be great. Farasa (talk) 21:15, 29 December 2007 (UTC)

• 
  lactose
• 
  sucrose
• 
  fructose

Look at the right most carbon in the two double ring systems. Lactose and fructose have hemiacetal groups and the ring structures can open up. The reducing sugar article does not really say "a reducing sugar is not connected with another sugar".....many disaccharides are reducing sugars. Note that the number 2 carbon of fructose links to the number 1 carbon of glucose in a sucrose molecule. Our figure for lactose has the two bridging carbons numbered (c-1 and c-4) and our figure for sucrose should also. --JWSchmidt (talk) 23:19, 29 December 2007 (UTC)