I do know that only the first link 'counts' as far as link juice is concerned.

Obviously, if you had 100 links pointing to the same page it would probably trigger an alert but for 3 links you should be fine.

Other than that, I would suggest having a single link to your pages.

I do know that only the first link 'counts' as far as link juice is concerned.

I've heard that as far as anchor text goes, but I'm not sure I'd buy into there being no link weight / PR impact ... It actually doesn't make sense (when talking about PageRank as the model for weighting) to not count subsequent links on a page as links to the destination, imo ... I can see eliminating the link text, because you prevent some 'stuffing' and 'possible manipulation' of 'topicality' on the receiving page, but if my goal is to determine the 'probability of the average surfer landing on the page' the subsequent links increase the probability of that happening and should increase the 'perceived importance' of the destination page relative to the other destination pages only linked once, and what PageRank tries to do is calculate the probability of an 'average surfer' landing on a given page (aka perceived importance relative to other pages), so, unless you've tested and KNOW they don't, imo, logic says subsequent links have to have an impact on PR.

I would say 'link evenly' if you want even distribution, or 'more heavily' to pages you want to flow a greater amount of PageRank to ... So, if you want more link weight passed to the page linked 3 times, it's likely you're getting it if you only have one link to other pages.

It's fine to have multiple links to the same destination on a single page.

Beyond that just try to make sure it all makes sense from a users point of view and you'll be fine.

I've heard that as far as anchor text goes

I was just about to comment the same. It's a lot harder to test fine gradations of PR vote than to test for anchor text influence.

And funny thing, at least one later test seemed to show that different anchor text can also work, if there's a unique fragment identifier (#fragmentname) at the end of the URL - even though the link is a repeat. This tilts the verdict even more toward what TMS was saying.

I've tested the PR passing of multiple links.

A second link doesn't count. It is ignored. It does not pass additional pagerank, nor does drop pagarank on the floor ala nofollow.

Like tedster says for anchor text influence, putting a unique #fragmentname on the second url may make it different in Google's eyes and cause it to pass pagerank. Its unclear to me if this will actually make it pass extra pagerank that can make the linked page rank better, or whether it will pass pagerank only to a section of the linked page so that Google can show the fragment in SERPs like they do to direct people deeper into Wikipedia articles upon occasion.

the home page does not have very much Page Rank, and I am trying to maximize the effectiveness of it

Today, PageRank is a more complex situation than when the original paper was published. One area to wrap your head around is what Google calls the "Reasonable Surfer" model. The original PR equations distributed PR through a "random surfer" model, but that simple approach is long gone. Here's a thread discussing Google's patent:

Google's "Reasonable Surfer" [webmasterworld.com]

Just as we know that external backlinks from a content area have more weight, you'll also find that internal links in your home page content area deliver more clout - so use them to good effect.

Well, no wonder it makes no sense to me:
The average surfer isn't reasonable; I was basing my conclusions on the wrong model! lol

---

Can I ask how you tested deadsea?

I really don't 'get' how they could use PR as their model for link scoring and then publishing their 'perceived importance' of the page to a visitor viewing the PR of a page and not count the second link on a page as an indication of 'more important' than pages with only one link, all other things being equal, of course.

I run experiments like the one I propose here:
[webmasterworld.com...]

I believe that the "reasonable surfer" model is not fully applied. I think that at this point the algorithm just discounts some links that would otherwise pass the full amount of pagerank. It appears that they apply a simple heuristic to identify such candidates. If the link is in the footer or is in a large list of links for example.

@tedster
@deadsea

And funny thing, at least one later test seemed to show that different anchor text can also work, if there's a unique fragment identifier (#fragmentname) at the end of the URL

putting a unique #fragmentname on the second url may make it different in Google's eyes and cause it to pass pagerank.

Just so I understand, you are talking about something like this:

One of the duplicate links from the homepage to the blue-widgets.html page should have a named anchor such as #bronze-age appended so that it would link to a specific area on the destination page?

So have ONE of the links as:

/blue-widgets.html#bronze-age

And on the destination /blue-widgets.html page, put a named anchor #bronze-age at the start of the paragraph where it talks about the history of blue widgets during the Bronze Age?

Or am I misinterpreting this?

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Also, let me simplify my original question a little bit and change it to a "What would you do?" type question.

Home page has a Tool Bar Page Rank of 2. It has 25 links to internal pages (no links from the home page point to external pages), and of those 25, only 20 are unique.

The left-hand navigation links are "higher" in the source than the text content links and the product photo links.

Next in the source code (and visually) come the content text links, then below them (in the code and visually) are the photo links.

So would you all move the most important keywords to the in content text links? The secondary keywords to the photo links? And the third level keywords to the left hand navigation links?

(This seems contrary to what a user would follow. Most visitors to the home page click the photo links to the products, followed by the left hand navigation links, followed by the in content text links. So while it might improve SEO, there is a good chance it will HURT conversions...)

In this case, pagerank only gets passed once. Anchor text only gets passed once, and I think it is Anchor 1, being first in the page source.
nav: <a href="pageB.html">Anchor 1</a>
content: <a href="pageB.html">Anchor 2</a>
photos: <a href="pageB.html">Anchor 3</a>


I've tried this in the past to try to get anchor 2 to be the anchor text that is passed. No idea if it works or not though.
nav: <a href="pageB.html" rel=nofollow>Anchor 1</a>
content: <a href="pageB.html">Anchor 2</a>
photos: <a href="pageB.html">Anchor 3</a>


This appears to pass pagerank and anchor text three times each. I'm sure Google would view it as black hat if you abuse it.
nav: <a href="pageB.html#foo">Anchor 1</a>
content: <a href="pageB.html#bar">Anchor 2</a>
photos: <a href="pageB.html#baz">Anchor 3</a>

Thank you, deadsea.