200,000 Elliptical Galaxies Point the Same Way

slashdot.org ^ | 29 aug 2007 | KJ Longo

[jbp1]

I have studied a sample of 200,000 elliptical galaxies with redshifts <0.20 from the Sloan Digital Sky Survey (SDSS) to investigate whether they tend to have their ellipticities aligned along a particular axis. The data show a 13 standard deviation signal for such an alignment. The axis is close to the spiral spin axis found previously and to that of the quadrupole and octopole moments in the WMAP microwave sky survey

[edsheppa]

The whole method is to find the great circle which maximizes the average ellipticity of galaxies in its vicinity,

No, that's not the method, you have misunderstood the procedure described in the paper. But at least now I understand what you're saying.

About those error bars. I don't think they mean what I was talking about. They don't behave properly. Assuming the 200K galaxies are evenly distributed across the whole sky, I'd expect the error bars to be proportional to 1/sqrt(sin(γ)) and they clearly aren't.

[dr_lew]

The paper states that the method is to find an orientation which maximizes the slope, s, of the best fit line, e = e0 + s * sin( gamma ), does it not?

But e0 is independent of orientation, so maximizing the slope is the same thing as maximizing the value e = e0 + s, at sin( gamma ) = 1, which is the average ellipticity at the equator of the maximizing orientation, as modeled by the equation.

What don’t I understand ?

[edsheppa]

so maximizing the slope is the same thing as maximizing the value e = e0 + s, at sin( gamma ) = 1

No. The slope is calculated least squares for each orientation over the (sin(γ),e) data points. The sin(γ)=1 bin is only one of the fifty for each orientation.

BTW, I was wrong about the magnitude of the statistical power of the sin(γ) bins. I calculated dV/dγ as proportional to sin(γ). I should have calculated dV/dsin(γ) and that's proportional to sin(γ)/cos(γ). The denominator increases the effect I mentioned quite a lot. For example, the ratio of the variances of the .9 to .1 bin would be ~20 instead of 9 and .9 to .2 would be ~10 instead of 4.5.

[dr_lew]

No. The slope is calculated least squares for each orientation over the (sin(γ),e) data points.

It's indisputable that the maximum value of s gives the maximum value of e0 + s. But you say "No" !

[SlowBoat407]

Would inertia be in effect for time travel? If so, we would keep moving in the same direction and speed (or in the case of going back in time, the reciprocal) while subject to the same gravitational forces, so in essence, we would stay in the same relative place.

Or not.

[edsheppa]

Oops, here is your remark to which I was actually objecting.

which is the average ellipticity at the equator of the maximizing orientation

[dr_lew]

That's a lot to include under an "oops" isn't it? And anyway, you edited out my qualification, " ... as modeled by the equation." which makes what I said perfectly true. In effect, you are ignoring my explanation here and reiterating your original objection.

The fact is that the authors chose to represent their conception by a linear correlation, and within this representation, and so within their conception, the equivalence I claim is valid. That is all.

[edsheppa]

That's a lot to include under an "oops" isn't it?

Maybe, I had just copied the wrong part of the sentence.

the equivalence I claim is valid

Here is your claim.

The whole method is to find the great circle which maximizes the average ellipticity of galaxies in its vicinity

and that is certainly consistent with what you said later

maximizing ... the average ellipticity at the equator of the maximizing orientation

assuming by "maximizing orientation" you mean the orientation that maximizes the average ellipticity at the equator.

But what I'm saying is that this is not what the author of that paper did. For each orientation he looked at the whole sky, not just the equator, and did a linear regression of ellipticity vs sin(γ). That can easily give a different result from simply choosing the orientation that maximizes ellipticity on its equator.

I'll be happy to give you a simple numerical example if you don't get what I'm saying.

[Lou L]

Quasars...

Try and ask somebody who doesnt believe in God to explain their existence.

Just curious--what kind of answer should I expect from someone who doesn't believe in God?

Being interested in astronomy since I was a little kid, I find that the presence of God is the only way I can believe in such things.

[dr_lew]

assuming by "maximizing orientation" you mean the orientation that maximizes the average ellipticity at the equator.

Each orientation is defined by an axis, and the "equator" is defined with respect to that axis. I believe I stated earlier that the equator is defined by sin(gamma) = 1, just as sin(gamma) = 0 defines the poles.

So again, the procedure is to find the orientation such that the fitted line has maximum slope, and this is the orientation that has the largest value of average ellipticity at the equator, as represented by the fitted line, and this value is e0 + s.

[edsheppa]

Because the line is fitted least squares, the average ellipticity at the equator may not be e0 + s.

[RightOnline]

"I read a book a few years ago that talked about a few concepts of superstrings’ interactions with matter (strings)."

Um..........ok..........

I read "Guitar Player" magazine. Lots of cool strings.

[kinoxi]

It predicted attractors comprised of branes surrounding the outside of the Universe. Pulling matter towards it. The alignment described in this article highly correlates. I thought it odd. Do you have a point?

[RightOnline]

Point? No. Just having fun with ya. Not all have a sense of humor, so.......never mind.

[null and void]

What’s brown and sounds like a bell?

[misanthrope]

Just curious, what do you guys do for a living?

[kinoxi]

Misunderstood, was getting late.

[Bigh4u2]

“What’s brown and sounds like a bell?”

Ok. I give.

What?

[null and void]

{{{{dung}}}}

[Old Professer]

Allah’s children, all.