Edwin Hubble and the Evidence for Universe Expansion

[Cerenkov]

[Moderator's note: Spin off from previous thread due to topic change.]

It's my understanding that Edwin Hubble used the Hooker telescope to measure the red shift of galaxies only within the Local Group of galaxies to determine that the universe was expanding.

As we see here... https://en.wikipedia.org/wiki/Local_Group

...the diameter of the Local Group is given as 3 megaparsecs. Some 97 megaparsecs short of the recession value cited by Ibix. Taking into account what Ibix said about random motion and given that the Hooker could only be used to make measurements within the Local Group, surely Hubble should not have detected an overall, universal expansion?

His measurements should have only detected the presence of random motion within the Local Group?

Even if my assumption (and that is all it is) that the Hooker could not be used to measure red shifts beyond the Local Group is incorrect, does that mean that Hubble measured the recessional movements of galaxies over 100 megaparsecs distant with that telescope?

Thanks for any help given to clear up my confusion.

Cerenkov.

 

[PeroK]

There's something bothering me about this discussion. Hopefully somebody can set me straight about it.

It's my understanding that Edwin Hubble used the Hooker telescope to measure the red shift of galaxies only within the Local Group of galaxies to determine that the universe was expanding.

As we see here... https://en.wikipedia.org/wiki/Local_Group

The thing that bothers me is that neither the word "red" nor the word "shift" appears on that page. Nor the words "expand" or "expansion".

Perhaps you could explain that?

 

[Cerenkov]

The thing that bothers me is that neither the word "red" nor the word "shift" appears on that page. Nor the words "expand" or "expansion".

Perhaps you could explain that?

That's my very point, PeroK.

If there's no significant red shift or expansion within the Local Group, then how did Hubble arrive at the conclusion that the universe was expanding?

The diagram posted by DaveC426913 is based upon Hubble's original, isn't it?

And it shows only galaxies within the Local Group, doesn't it?

And none of the data points are from more than 2 megaparsecs?

So, how could Hubble draw the conclusion of galactic recession when Ibix says that this effect is only seen at distances beyond 100 megaparsecs?

I wasn't aware that the Hooker telescope was able to measure galactic recession at those distances.Cerenkov.

 

[PeroK]

That's my very point, PeroK.

If there's no significant red shift or expansion within the Local Group, then how did Hubble arrive at the conclusion that the universe was expanding?

By looking at more distant galaxies! Where there was red shift.

 

[Bandersnatch]

By looking at more distant galaxies! Where there was red shift.

Hubble's original data doesn't go past roughly 2 Mpc, is the point.

 

[Cerenkov]

Exactly Bandersnatch.

So, if only galaxies beyond 100 megaparsecs display this recession and if Hubble's data extends out to 2 megaparsecs, what gives?

How could he arrive at the conclusion that the universe was expanding?

His data set falls short of the critical threshold by 98 megaparsecs.

Help!

 

[PeroK]

Hubble's original data doesn't go past roughly 2 Mpc, is the point.

So, what is the answer?

 

[Bandersnatch]

Ok, so I'm not terribly clear on that, but let me give you my general intuition that may or may not be the full answer.
There's two bits at play. One, less important, is that the original data was messy and not that well calibrated, so that some distances are underestimated by a factor of 2 or more. That still is way short of 100 Mpc, though.
Which brings me to the second point - what the 100 Mpc ballpark tells you is not as much the distance at which cosmological redshift first starts to show, but the distance at which the expansion dominates the dynamics. It's where the recessional velocities can be expected to be large enough to exceed escape velocities of the masses of galaxies contained within that radius. That is to say, for most of even nearby galaxies the recessional motion set in by the expansion shows quite well, but they are still gravitationally bound. There was simply not enough time in the history of the universe to allow these galaxies to sufficiently slow down or reverse motion.
In the vicinity of very massive, very close members of various groups this time scale is short enough, so that after 13+ Gy we catch them already orbiting/colliding/merging. But the more distant, smaller clusters are still on the outbound legs of their first ever orbits, so to speak.

 

[PeroK]

So, if only galaxies beyond 100 megaparsecs display this recession and if Hubble's data extends out to 2 megaparsecs, what gives?

How could he arrive at the conclusion that the universe was expanding?

His data set falls short of the critical threshold by 98 megaparsecs.

Help!

You're right. It's not clear how Hubble came to a definite conclusion about expansion. I found this:

https://www.pnas.org/content/101/1/8

"Although his distances had serious errors due to confusing two types of Cepheids, and blurring bright gas clouds with bright stars, in 1929, Hubble was able to sort nearby galaxies from distant ones well enough not to miss the connection between distance and velocity."

 

[Cerenkov]

Hmmm...

It would be very satisfying to know just how Hubble made the connection between distance and velocity, given that his dataset (and the capability of his telescope?) falls so far short of the distance where these things become readily apparent.

As a non-scientist limited to Basic-level topics I simply cannot provide any answers here.

All I can do is to draw attention to this mismatch and ask for help.

Thank you.

Cerenkov.

 

[Cerenkov]

Oh dear!

Things seem to have become even more complicated.

PeroK's link says this.

https://www.pnas.org/content/101/1/8

The program of measuring galaxy spectra had been initiated a decade earlier by Vesto Melvin Slipher at the Lowell Observatory in Arizona. By 1923, after heroic efforts with small telescopes and slow spectrographs, Slipher had compiled a list of velocities for 41 galaxies, 36 of which were receding from us, and the largest of which was moving away at 1,800 km/s. This intriguing list was published in Arthur Stanley Eddington's textbook on general relativity, The Mathematical Theory of Relativity. Hubble cites no source for the radial velocities in table 1 of ref. 1, except for the four new ones from his Mount Wilson colleague, Milton Humason, but every one of the galaxies is one of Slipher's, and the list of velocities is almost identical to that in Eddington's book.

So, if every galaxy used by Hubble comes from Slipher's work, this was done using the Lowell Observatory telescope, where Slipher spent his entire career.

https://en.wikipedia.org/wiki/Lowell_Observatory

All of the telescopes Slipher might have used there are significantly smaller than the 100 inch Hooker, used by Hubble. Which suggest to me that their light-grasp would be inferior.

Now I'm REALLY confused.

Cerenkov.

 

[Bandersnatch]

falls so far short of the distance where these things become readily apparent.

But they are apparent. Your confusion seems to hinge on this bit:

Recession of galaxies is only seen at large distances, over 100Mpc, if memory serves.

Which is overstating the issue. As was explained in post #20, that's not where the recession first shows up (see modern graph below) - that's the ballpark where recession velocities can be universally thought to be larger than escape velocities (and I'm not sure how married we should be to this number anyway).
I.e. if you waited infinite amount of time, the aggregations of galaxies presently below that scale would not move to infinity but remain bound. Which is very much not the same statement as saying they're not receding now. For most, the initial impulse is still there, and shows in the data.

Hubble's original paper for reference: link

Modern data shows clearly that the law holds below 100 Mpc:

 

[PeterDonis]

There's something bothering me about this discussion.

I'm not sure what would bother you about the fact that 90 years ago our evidence for the expansion of the universe was sparser and had a lower level of confidence than today. We've collected a lot more evidence since then.

 

[Cerenkov]

Peter and Bandersnatch,

Of course I do not dispute or doubt the evidence that has accumulated in the intervening 90 years since Hubble used Slipher's data.

What I find bothersome is understanding how Hubble could have drawn his conclusion, given the following.

1. That Slipher used telescopes with an inferior light-gathering ability than the 100 inch Hooker telescope.

2. That this dataset was sparse and at a lower level of confidence than today.

3. That the dataset included only galaxies out to 2 mpc.

4. That even if we slash the distance at which recession becomes apparent down to 10 mpc, that's still five times further than any of the Slipher dataset Hubble used to draw his conclusion.

I just can't see how Hubble could have done it.

That said, if it's ruled that tackling my confusion lies outside the remit of what can be discussed in this thread, then I will, of course, abide by that ruling and keep my confusion to myself.

Thank you.

Cerenkov.

 

[Bandersnatch]

4. That even if we slash the distance at which recession becomes apparent down to 10 mpc,

But this is still just as untrue as with 100 Mpc. Check the original paper - the velocity/distance recessional trend is visible even with the 2 Mpc data he had.

 

[Cerenkov]

But this is still just as untrue as with 100 Mpc. Check the original paper - the velocity/distance recessional trend is visible even with the 2 Mpc data he had.

I see it Bandersnatch.

So, perhaps I'm being confused by Ibix's comments? Specifically, this...

Essentially, all galaxies have some random motion on top of the uniform expansion, and for galaxies nearer than that the typical magnitude of the random motion is comparable or greater than that of expansion.

Which I have taken to mean that the random motion of galaxies within the Local Group (up to 2 mpc) is comparable or greater than the value of their recessional velocity.

Which caused me to wonder how Hubble could have drawn the conclusion of universal recession.

Are you suggesting that I disregard what Ibix says in favour of what Hubble graph indicates?Cerenkov.

 

[Ibix]

Are you suggesting that I disregard what Ibix says in favour of what Hubble graph indicates?

Well it rather looks like the evidence is that I was making statements that were too strong, so I'd say disregard me...

 

[Cerenkov]

Ah thanks Ibix.

I am now satisfied that I understand this better and than my confusion is at an end.

My thanks to PeterDonis, Bandersnatch, PeroK and Ibix for their help.

Cerenkov.

 

[Bandersnatch]

Yes, that statement was too strong. It's definitely true for many members of the LG (there's plenty of straight-up satellites of MW, Andromeda or M33), but not for others.

Can't be blaming Ibix too much, though. I've been guilty of throwing similar statements around in the past, I'm sure.

 

[Ibix]

So, what I originally wrote was:

Recession of galaxies is only seen at large distances, over 100Mpc, if memory serves. Essentially, all galaxies have some random motion on top of the uniform expansion, and for galaxies nearer than that the typical magnitude of the random motion is comparable or greater than that of expansion. So nearby galaxies (like Andromeda, not even 1Mpc distant) can be moving away from us, towards us, or purely transversely. Uniform expansion becomes the dominant effect beyond that.

But what I should say is more like: Hubble's Law applies to the average velocity of galaxies at a particular distance from us. All galaxies have some (essentially) random motion on top of their Hubble average motion, and where the Hubble velocity is small it's not uncommon for the random velocity component to exceed the average - so we see quite a few nearby galaxies moving towards us.

 

[JimJCW]

So, if only galaxies beyond 100 megaparsecs display this recession and if Hubble's data extends out to 2 megaparsecs, what gives?

How could he arrive at the conclusion that the universe was expanding?

His data set falls short of the critical threshold by 98 megaparsecs.

Help!

Sorrell’s 2009 paper, Misconceptions about the Hubble recession law, might be interesting. Here is part of the abstract of his paper:

Almost all astronomers now believe that the Hubble recession law was directly inferred from astronomical observations. It turns out that this common belief is completely false. Those models advocating the idea of an expanding universe are ill-founded on observational grounds. This means that the Hubble recession law is really a working hypothesis.​

Shown below is a paragraph from the Introduction section of the paper:

The purpose of the present study is to ask and answer a radical question that is seldom asked. The question is this: Do astronomical observations necessarily support the idea of an expanding universe? Almost all cosmologists believe as sacrosanct that the Hubble recession law was directly inferred from astronomical observations. As this belief might be ill-founded, it is necessary to critically assess the Hubble law; and highlight the principal assumption that led Hubble to deduce his law for the recession motion of galaxies.​

 

[PeroK]

Sorrell’s 2009 paper, Misconceptions about the Hubble recession law, might be interesting. Here is part of the abstract of his paper:

Almost all astronomers now believe that the Hubble recession law was directly inferred from astronomical observations. It turns out that this common belief is completely false. Those models advocating the idea of an expanding universe are ill-founded on observational grounds. This means that the Hubble recession law is really a working hypothesis.​

Shown below is a paragraph from the Introduction section of the paper:

The purpose of the present study is to ask and answer a radical question that is seldom asked. The question is this: Do astronomical observations necessarily support the idea of an expanding universe? Almost all cosmologists believe as sacrosanct that the Hubble recession law was directly inferred from astronomical observations. As this belief might be ill-founded, it is necessary to critically assess the Hubble law; and highlight the principal assumption that led Hubble to deduce his law for the recession motion of galaxies.​

There's a world of difference between asking how Hubble managed to infer the recession law from what looks like very limited data in 1929 and denying the extensive observations of the last hundred years.

Sorrell wheels out the tired light hypothesis and I suggest this marks him out as something of a crackpot.

 

[weirdoguy]

Sorrell wheels out the tired light hypothesis

And so do you @JimJCW every now and then, even though you've been repeatedly told what's the consensus amongst mainstream scientists about both BB theory and dead tired light model.

 

[russ_watters]

Which I have taken to mean that the random motion of galaxies within the Local Group (up to 2 mpc) is comparable or greater than the value of their recessional velocity.

Which caused me to wonder how Hubble could have drawn the conclusion of universal recession.

Perhaps the discussion has moved past this, but I think there is a common misconception at work here that the signal has to be larger than the noise to be detectable. It doesn't. Since the noise is by nature random, if you add enough data points it averages itself toward zero and the signal becomes apparent. You can see it in the graph.

 

[Cerenkov]

Perhaps the discussion has moved past this, but I think there is a common misconception at work here that the signal has to be larger than the noise to be detectable. It doesn't. Since the noise is by nature random, if you add enough data points it averages itself toward zero and the signal becomes apparent. You can see it in the graph.

Yes, the discussion has moved past this, Russ...

But your point is well made and welcome to me.

I didn't understand this and now I have something new to chew over, digest and assimilate.

Thank you very much.

Cerenkov.

 

[InkTide]

if you add enough data points it averages itself toward zero

I believe the noise itself averages to whatever the mean of the noise is, so the error in that average tends to zero - then you can examine the underlying signal. This is of course provided the noise is truly random (i.e. Gaussian).

http://www.dspguide.com/ch2/3.htm

 

[PeroK]

Ah thanks Ibix.

I am now satisfied that I understand this better and than my confusion is at an end.

My thanks to PeterDonis, Bandersnatch, PeroK and Ibix for their help.

Cerenkov.

I've been reading The First Three Minutes, by Steven Weinberg and he says:

Actually, a look at Hubble's data leaves me perplexed how he could reach such a conclusion - galactic velocities seem almost uncorrelated with their distance. In fact, we would not expect any neat correlation of proportionality between velocity and distance for these 18 galaxies - they are all much too close. It is difficult to avoid the conclusion that ... Hubble knew the answer he wanted to get.

However that may be, by 1931 the evidence was greatly improved ...

 

[Buzz Bloom]

I found the investigation in this thread to be quite interesting about Hubble's data leading to the expansion concept with values. I am curious about what a diagram (similar to the modern diagram presented in post #12 of velocity vs. distance) would look like of the limited data that Hubble used. Does anyone know where such a diagram might exist?

 

[JimJCW]

I found the investigation in this thread to be quite interesting about Hubble's data leading to the expansion concept with values. I am curious about what a diagram (similar to the modern diagram presented in post #12 of velocity vs. distance) would look like of the limited data that Hubble used. Does anyone know where such a diagram might exist?

How do I get to the ‘velocity vs. distance’ thread?

 

[Buzz Bloom]

Hi @JimJCW:

I would like to be able to answer your question in the previous post, but I do not understand what your question means. Please explain.

Regards,
Buzz

 

[JimJCW]

Hi @JimJCW:

I would like to be able to answer your question in the previous post, but I do not understand what your question means. Please explain.

Regards,
Buzz

Where can I see

post #12 of velocity vs. distance?​

​

Please provide the link to that thread.

 

[JimJCW]

Hi @JimJCW:

I would like to be able to answer your question in the previous post, but I do not understand what your question means. Please explain.

Regards,
Buzz

Are you talking about the current thread,

Edwin Hubble and the Evidence for Universe Expansion?​

 

[Bandersnatch]

Where can I see

post #12 of velocity vs. distance?​

​

Please provide the link to that thread.

It's the 12th post (which happens to be mine) in this very thread. Each post in any thread is numbered, the assigned number being displayed in the top-right corner of said post.

I found the investigation in this thread to be quite interesting about Hubble's data leading to the expansion concept with values. I am curious about what a diagram (similar to the modern diagram presented in post #12 of velocity vs. distance) would look like of the limited data that Hubble used. Does anyone know where such a diagram might exist?

You just have to look at Hubble's original diagram.
Fig. 1 on the fifth page here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC522427/
If you want the same set of objects with modern distance/velocity values, you can plot it yourself. Use the tables in that paper to find each object in any public database.

 

[JimJCW]

It's the 12th post (which happens to be mine) in this very thread. Each post in any thread is numbered, the assigned number being displayed in the top-right corner of said post.
You just have to look at Hubble's original diagram.
Fig. 1 on the fifth page here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC522427/
If you want the same set of objects with modern distance/velocity values, you can plot it yourself. Use the tables in that paper to find each object in any public database.

Thanks! I thought there was a thread named ‘velocity vs. distance’ somewhere.

 

[Buzz Bloom]

Are you talking about the current thread,

Yes.