LGM biotypes and haplogroups

[Karhunkynsi]

I posted these paleovegetation maps some time ago. It's interesting how different the routes for paternal hg's start to look when they are put into "Ice Age"-context.

Here's my educated guess:

1) N1's were roaming the biotype that is called Mammoth Steppe. This biotype spanned from France to Beringia. N1's likely inhabiting rather the eastern part and travelling towards west.

2) R1's were strictly in South Asia, west of South Asia was badlands, so their dispersal towards west must have happened after LGM.

3) E's were all landlocked into Subsaharan Africa. Megadesert kept them isolated from the Eurasia.

4) I's were at South Europe.

5) J's, pure guess.

6) O's were at Southeast Asia, where they started to expand towards north.

7) Q's were roaming the same Mammoth Steppe as N1's, but were concentraded closer to Beringia (for obvious reason).

What do you think ? Lot of HG's missing as I dont know have that clear opinion about those.

[Aware_Dog]

LGM was ~ 25-13KYA, during that time E had already traveled up to Northern Africa through the Nile Corridor, Per the TMRCA of E1b1b and E1b1b1a.

[Karhunkynsi]

LGM was ~ 25-13KYA, during that time E had already traveled up to Northern Africa through the Nile Corridor, Per the TMRCA of E1b1b and E1b1b1a.

The map represent situation roughly 18kya ago. I doubt there was any E1b1b* at North Africa back then.

[Aware_Dog]

The map represent situation roughly 18kya ago. I doubt there was any E1b1b* at North Africa back then.

Cruciani et. al 2007:

E-M78 belongs to clade E3b (E-M215). On the basis of robust phylogeographic considerations, an eastern African origin has been proposed for E-M215 (Underhill et al. 2001; Cruciani et al. 2004), with a coalescence time of 22.4 ky (95% C.I. 20.9-23.9 ky; recalculated from Cruciani et al. 2004, see Materials and Methods). A north-eastern African origin for haplogroup E-M78 implies that E-M215 chromosomes were introduced in north-eastern Africa from eastern Africa in the Upper Paleolithic, between 23.9 ky ago (the upper bound for E-M215 TMRCA in eastern Africa) and 17.3 ky ago (the lower bound for E-M78 TMRCA here estimated, fig. 1). In turn, the presence of EM78 chromosomes in eastern Africa can be only explained through a back migration of chromosomes that had acquired the M78 mutation in north-eastern Africa.

E1b1b*=E3b*=E-M215

[Bandar Qasim]

3) E's were all landlocked into Subsaharan Africa. Megadesert kept them isolated from the Eurasia.


The Nile river has always formed a corridor between the Mediterranean and (Sub Saharan) East Africa.

There never were natural barriers between those two regions preventing migration.

[beyoku]

Cruciani et. al 2007:


E1b1b*=E3b*=E-M215

I think you guys are somewhat both correct. One problem is Cruciani's data (i believe) only deals strictly with the genetics so it doesn't take into account where ice may have been 1000's of Years ago. I am sure we can find other genetic articles that talk about a certain Haplogroup origin in Europe at whatever time, even though in reality it may have been covered in a mile high ice sheet.

That being said Battaglia et al 2008 article DOES use some geological/archeological data in its conclusion stating
a Refugium existed on the border of present-day Sudan and Egypt, near Lake Nubia, until the onset of a humid phase around 8500 BC Noting these lineages were confined to the South and pushed northward much recently. I am pretty sure that it is no coincidence that most Modern Africans and much of the ancient skeletal material found in Africa is called "Tropically adapted". The refugium Battaglia speaks of basally starts at the Tropic of Cancer....Again showing that for the most part Africans adapted to an area between souther Egypt (Tropic of Cancer) and South Africa (Tropic of Capricorn)

Your map and hypothesis most definitely makes sense when looking at the long term isolation of Pn2 Sub Clades in East Africa. Also while E-M78* is very rare it is still found in Egypt as well as SOUTHERN Sudan.....which is usually overlooked. I think this also explains the Push of E1b1a primarily to the south only.

I have not done enuf specific research to know how much the mega arid desert affected the mega lakes that were in the Sahara. Therefore I cannot tell you what type of exceptions to the rule their may have been : Wadi Halfa, Holocene Lake Mega Chad, Ancient Lakes of the Western desert of Egypt, Wadi Al-Hayat, Gebel Uwainat, Napta Playa, etc.

2 cents

[Karhunkynsi]

Here is more detailed map of Africa, key is the same as in OP. There is savannahs, rain forest and grasslands at Central West Africa. That looks like place where humans were, they were not at arid mega desert.

Nile connecting SSA-region to Mediterranean was a good point. It seems that Nile actually did exist allready during LGM, it was smaller but it was there. I dont think anyone was using it at that time tho' (as route, for fishing sure). There needs to be something to pull people, Nile was flowing nice and quiet in the middle of the dry sandy desert basically. Why would anyone travel that route ? And what for ?

There is neat little slot of woodlands in the vicinity of the horn. That could be the place where E-M78 mutation happened. Thats also pretty close to (White) Nile headwaters. Once the climate got more moist, animals expanded to "savannah-Sahara", then people would follow.

[Karhunkynsi]

I know we have few J* specialists here, please put the J*'s into the map. I was simpy guessing but looking at that terrain type it is highly unlikely that J*'s were where I put them. Where were they ?

[Bandar Qasim]

And another thing, you have put E in the Congo forest. But that area probably had very few E's during the LGM and was mostly A's and B's (Khoisanid-Pygmies). Bantu migration carrying E3a into that region happened much later.

[Karhunkynsi]

And another thing, you have put E in the Congo forest. But that area probably had very few E's during the LGM and was mostly A's and B's (Khoisanid-Pygmies). Bantu migration carrying E3a into that region happened much later.

True, thanks for pointing that out.

I renew my question about J*'s , any taker ?

Another one of my interests is missing G*'s.

[Jaska]

1) N1's were roaming the biotype that is called Mammoth Steppe. This biotype spanned from France to Beringia. N1's likely inhabiting rather the eastern part and travelling towards west.
Interesting, but what is the basis for this educated guess? I mean, N1 in France 18 000 years ago...

[Karhunkynsi]

Interesting, but what is the basis for this educated guess? I mean, N1 in France 18 000 years ago...

Not in France no but on the same biotype, which spanned from France to Beringia. The Mammoth Steppe (http://mammothsteppe.com/index.php?option=com_content&view=article&id=8&Itemid=14).



During the Late Pleistocene the environment of the northern territories of Eurasia, ranging from Western Europe to Alaska, was generally open, tending toward steppic. It is known as the Mammoth Steppe, a biotope remarkable for its unusual combination of mammal species when compared to the present-day distribution. The ecosystem of the Mammoth Steppe collapsed during the period that marks the Pleistocene – Holocene transition and was replaced by the modern tundra, taiga, and steppe belts of Eurasia.


There may have been some occasional N1-expeditions as west as France but it must have been rare occasions indeed. Majority of the fauna, mammoths and other cudly little furries, lived where is now fareast Europe and all the way to Beringia in the east. According to Mirabel et al 2009, mutation for N1b (N-P43) occurred somewhere in East Europe roughly 14kya ago. So N1 must have been present in this biotype.

According to Rootsi et al 2006 mutation for N1 (back then she thought that it was N, but after discovery of new snp's it was actually N1) arose on the eastern end of this biotype roughly 20kya ago.

Mobile hunters carrying mutation for N1 roamed this Mammoth Steppe, as did hg Q carrying future Amerindians and Siberians.

Imo, looking at those maps is very revealing. It really brings to life the situation back then and so to speak, adds flesh to the bones how and where humans dwelled.

[Lemminkäinen]

Here is more detailed map of Africa, key is the same as in OP. There is savannahs, rain forest and grasslands at Central West Africa. That looks like place where humans were, they were not at arid mega desert.

Nile connecting SSA-region to Mediterranean was a good point. It seems that Nile actually did exist allready during LGM, it was smaller but it was there. I dont think anyone was using it at that time tho' (as route, for fishing sure). There needs to be something to pull people, Nile was flowing nice and quiet in the middle of the dry sandy desert basically. Why would anyone travel that route ? And what for ?

There is neat little slot of woodlands in the vicinity of the horn. That could be the place where E-M78 mutation happened. Thats also pretty close to (White) Nile headwaters. Once the climate got more moist, animals expanded to "savannah-Sahara", then people would follow.

Sahara could have been a fertile place about 10000 years ago, perhaps also earlier.

http://www.time.com/time/magazine/article/0,9171,808593,00.html




.

[Jaska]

There may have been some occasional N1-expeditions as west as France but it must have been rare occasions indeed. Majority of the fauna, mammoths and other cudly little furries, lived where is now fareast Europe and all the way to Beringia in the east. According to Mirabel et al 2009, mutation for N1b (N-P43) occurred somewhere in East Europe roughly 14kya ago. So N1 must have been present in this biotype.
They made a bad mistake there, when they did not separate N1b-A and N1b-E in those calculations. N1b-E is only one subgroup of N1b-A, and the diversity must be calculated by lineage, not just by region. This error was pointed out by Malyarchuk and Derenko (from page 6 on):

http://www.mv.helsinki.fi/home/jphakkin/N1b.pdf

Consequently, there is no basis for the claim that N1b was born in Europe. Also N1c and N1a seem to have been born in Asia, so it is improbable that there were many N1's in Europe. Of course it is possible, and the dating and descenting of European N1 may tell if there were N1 in Europe already so early. At this point I see it improbable.

[Karhunkynsi]

Sahara could have been a fertile place about 10000 years ago, perhaps also earlier.


That is the period I associate with spread of E1b1b* out of Africa all the way to mediterranean and Balkans (and West Asia obviously).

---------- Post added 2010-12-31 at 18:18 ----------



Consequently, there is no basis for the claim that N1b was born in Europe. Also N1c and N1a seem to have been born in Asia, so it is improbable that there were many N1's in Europe. Of course it is possible, and the dating and descenting of European N1 may tell if there were N1 in Europe already so early. At this point I see it improbable.

The thing here is: There really was no Europe. There was the Mammoth Steppe.

[Jaska]

The thing here is: There really was no Europe. There was the Mammoth Steppe.
Interesting, but we cannot assume that just one population owned all the Mammoth Steppe. If N1 spread from the east, it met some other people in the west. If they "mixed" with those western peoples, we should see traces of this old N1 in central or Western Europe. If we don't have so old traces, we must conclude that N1 spread to the west only later.

[jibarodepr]

True, thanks for pointing that out.

I renew my question about J*'s , any taker ?

Another one of my interests is missing G*'s.
G was in South Europe, the Middle east and West Asia.

J is on the Middle East, South Europe, West Asia and North Africa(Maghreb) throguht its sub-clades J1 and J2.

[Karhunkynsi]

G was in South Europe, the Middle east and West Asia.

J is on the Middle East, South Europe, West Asia and North Africa(Maghreb) throguht its sub-clades J1 and J2.

It is very unlikely that they (or anyone else) were in Middle East. Please take a look at those maps in OP. I doubt there was anyone at Middle East at that time frame (LGM). What does West Asia mean ? Isnt it just another term for Middle East anyway ?

Possible locations are really South Europe + Anatolia, Mammoth Steppe (especially terrain type 16 imo) or South Asia.

---------- Post added 2011-01-01 at 21:49 ----------


If they "mixed" with those western peoples, we should see traces of this old N1 in central or Western Europe.


Not really. There was very few people back then. Atm it seems like there is practically nothing left of those paleolithic inhabitants of Central or Western Europe. R1-decendants form the absolute majority pretty much everywhere and they were locked into South Asia still in LGM. Even the neolithic "farmer-mtdna" has barely survived till modern day (R1b* is thriwing tho' and it seems to be neolithic/bronze age arrival at West Europe).



If we don't have so old traces, we must conclude that N1 spread to the west only later.

I think the major hunting grounds of N1's were at Southwestern shore of paleolake-Mansi (the big lake at central Eurasia, just south of glaziers).

My reasoning behind this is that the steppic penetration into glaziers/polar desert, east of Mansi-lake, opened and allowed connection to Beringia before the actual LGM ended.

If N1's would have been on eastern side of lake Mansi, they would have ended up in Americas, just like Q's did. They couldnt expand to that direction until later, when the Q's had allready passed to Americas and Bering route was closed.

It's also very convinient location if looking the N1-decendant major distribution pattern today. The North Eurasia from Norway to Beringia.

[Jaska]

Not really. There was very few people back then. Atm it seems like there is practically nothing left of those paleolithic inhabitants of Central or Western Europe. R1-decendants form the absolute majority pretty much everywhere and they were locked into South Asia still in LGM. Even the neolithic "farmer-mtdna" has barely survived till modern day (R1b* is thriwing tho' and it seems to be neolithic/bronze age arrival at West Europe).
You may be right: we may lack all the traces of pre-Neolithic Europeans. Although in maternal lineages the datings are older than in paternal lineages. How about some rare but widespread paternal haplogroup like F*?


I think the major hunting grounds of N1's were at Southwestern shore of paleolake-Mansi (the big lake at central Eurasia, just south of glaziers).

My reasoning behind this is that the steppic penetration into glaziers/polar desert, east of Mansi-lake, opened and allowed connection to Beringia before the actual LGM ended.

If N1's would have been on eastern side of lake Mansi, they would have ended up in Americas, just like Q's did. They couldnt expand to that direction until later, when the Q's had allready passed to Americas and Bering route was closed.

It's also very convinient location if looking the N1-decendant major distribution pattern today. The North Eurasia from Norway to Beringia.
Sounds credible.

[Paul_Johnsen]

Not in France no but on the same biotype, which spanned from France to Beringia. The Mammoth Steppe (http://mammothsteppe.com/index.php?option=com_content&view=article&id=8&Itemid=14).



There may have been some occasional N1-expeditions as west as France but it must have been rare occasions indeed. Majority of the fauna, mammoths and other cudly little furries, lived where is now fareast Europe and all the way to Beringia in the east. According to Mirabel et al 2009, mutation for N1b (N-P43) occurred somewhere in East Europe roughly 14kya ago. So N1 must have been present in this biotype.

According to Rootsi et al 2006 mutation for N1 (back then she thought that it was N, but after discovery of new snp's it was actually N1) arose on the eastern end of this biotype roughly 20kya ago.

Mobile hunters carrying mutation for N1 roamed this Mammoth Steppe, as did hg Q carrying future Amerindians and Siberians.

Imo, looking at those maps is very revealing. It really brings to life the situation back then and so to speak, adds flesh to the bones how and where humans dwelled.

Rootsi, Mirabel and Derenko all use "Zhivotovsky-rates". I think father-son-rates are better. If "Zhivotovsky-rates" are true, this would mean that a TMRCA for R1a and R1b would be around 60ky ago (and that R1b expanded in Western Europe about 12 ky ago).

[Karhunkynsi]

How about some rare but widespread paternal haplogroup like F*?


Another usual suspect would be T* (formerly known as K2*).

---------- Post added 2011-01-02 at 16:05 ----------


Rootsi, Mirabel and Derenko all use "Zhivotovsky-rates". I think father-son-rates are better. If "Zhivotovsky-rates" are true, this would mean that a TMRCA for R1a and R1b would be around 60ky ago (and that R1b expanded in Western Europe about 12 ky ago).

Tbh I somewhat doubt all of these different calculation methods. It sounds allmost impossible to create formula which could correct the different population sizes and expansion rates of different haplogroups.

I mean it is obvious that if you have plenty of men belonging to same haplogroup you have much more opportunities for mutations, while those hg's that have only few carriers have much less chances for mutations. It's all about odds.

Another aspect, which affects what I wrote earlier is oviously the culture. Kyrgyz are something like 70% R1a1*, Yakut are similary extremely high for N1c*, yet they both have very small diversity due to their strict patrilineal nomadic background.

So not only should the formula take account of different population sizes, different expansion speeds but also cultural phenomenoms. I dont think that will ever be possible. What fits one hg is not necessary the solution for another.

Anyhow, there are more maps here, ranging from 18kya, 8 kya to 5 kya ago.:

http://www.esd.ornl.gov/projects/qen/adams4.html

8kya ago the vegetation corridor from South Asia to Anatolia+Europe was fully open (no more badlands). So atleast then it was possible to migrate out from South Asia.

As those chances happend slow, the 8 kya most be considered as too late for the earliest oppurtunity to expand westwards. The Opportunity may have existed allready 12kya ago, which could infact allow R1's out of South Asia, perhaps into Anatolia or Iran ? Afaik there is hardly any R1b* at India so the mutation happened rather after some of the R1's trecked westwards ?

Also Sahara had turned into this 8kya ago:

North Sahara:


5. Tropical semi-desert (sparse scrub or sparse grassland)

(Corresponds to subdivision of Olson desert and semi-desert)

Less than 2% vegetation cover above 80cm off the ground. 25-4% vegetation cover between 0 and 80cm off the ground, during an average year.


And South Sahara:



6. Tropical grassland (fairly closed grassland without many trees or shrubs)

(Corresponds to subdivision of Olson tropical savanna and woodland, and overlaps with warm or hot shrub and grassland)


I think the R1b*'s migrated into this biotype, especially to the biotype 6, (via Nile walley) and got landlocked into 'Cameroon' after the Sahara got worse again.

E1b1b's expanded to exactly opposite direction, north, gave high five to R1b's on the way and settled as far as Levant, where they then begun expanding further as the first farmers. Again the 8kya is prolly too late and the migrations happened few thousend years earlier. Atleast for E1b1b's.

[Paul_Johnsen]

Another usual suspect would be T* (formerly known as K2*).

---------- Post added 2011-01-02 at 16:05 ----------



Tbh I somewhat doubt all of these different calculation methods. It sounds allmost impossible to create formula which could correct the different population sizes and expansion rates of different haplogroups.

I mean it is obvious that if you have plenty of men belonging to same haplogroup you have much more opportunities for mutations, while those hg's that have only few carriers have much less chances for mutations. It's all about odds.

Another aspect, which affects what I wrote earlier is oviously the culture. Kyrgyz are something like 70% R1a1*, Yakut are similary extremely high for N1c*, yet they both have very small diversity due to their strict patrilineal nomadic background.

So not only should the formula take account of different population sizes, different expansion speeds but also cultural phenomenoms. I dont think that will ever be possible. What fits one hg is not necessary the solution for another.


There is no reason why demographics should influence the mutation rate. There is no way that a fathers genes can possible that his son has or doesn't have any brothers when he is being born (or after his son is born). The Kyrgyzstan R1a and the Yakut N1c both have very low diversity and therefore low TMRCA. There was a major R1a-founder and a major N1c-founder in the respective populations very recently. I don't see any reason not to use the father-son-rates to estimate when this person lived. The argument is that these founders entered the population 3,5 times earlier as a law of nature, I find difficult to take seriously.

[Karhunkynsi]

There is no reason why demographics should influence the mutation rate.


Ofcource it does! Mutations occure randomly. If we have two groups of men, mainly carrying same HG. Other group contains 100 men and other group 100.000 men. I'm pretty damn sure that the group containing 100.000 men has more mutations inside the "test period". This is because mutations occure randomly. More men, more chances for random mutation events.

Perhaps I didnt write clear enough in my earlier post. I had two arguments. 1) Culture shapes y-dna pool of population b) Large population size increases mutations in population.

I'm not defending Zhivotovsky's method any more than I'm arguing against pedigree-method. I think neither of them is very good. I also dont think there ever will be any bullet proof system because there will never be algorithm that could could handle all epigenetic (cultural) parameters and autocorrect the bias caused by different population sizes and histories.

I used Yakuts and Kyrgyz as examples as they are relatively well studied populations. Well, Yakuts are throughout studied. Their y-dna pool clearly reflects their culture, which is strictly patrilineal nomadism. Patrilineal clans are formed so that all male members of the clan are decendants of the same founding patriarch. This is clearly visible in Yakut ydna pool. Culture shapes populations genes.

[Jaska]

Here I agree with Karhunkynsi.

Both methods have resulted in "right" estimates, and both "wrong", which only tells that demographic etc. processes cannot be taken into account.

One weakness is common to all these calculative methods: if I got it right, they cannot find the inner structure of haplogroup/cluster, but they only rely on the variation/diversity/variance etc. Only reconstructive method can give more credible results.

For example, in the case of 100 men agains 10 000 men, it is most important to reconstruct the founder haplotype (FHT) of the cluster (by comparing to other clusters of the haplogroup and parallel haplogroups). After this has been done, we can see that in both groups there is a star-like structure, and all the haplotypes are quite close to the FHT of the cluster in terms of GD. Then, no matter how many men there are, we get the [B]depth of diversity by the means of reconstructive method. As a result, the age of both clusters is quite similar.

On the other hand, by the means of calculative methods we only get width of diversity (alleles per locus etc.), and at this level it is clear that the age of the cluster of 10 000 men is much older than the age of the cluster of 100 men, because there are much more variation in the cluster of 10 000 men.

[Paul_Johnsen]

Ofcource it does! Mutations occure randomly. If we have two groups of men, mainly carrying same HG. Other group contains 100 men and other group 100.000 men. I'm pretty damn sure that the group containing 100.000 men has more mutations inside the "test period". This is because mutations occure randomly. More men, more chances for random mutation events.

How can mutation rates be influenced by demographics? what is the underlying force at work? how does the cell know that the individual has or doesn't have living relatives?

Mutation rates are assumed to happen at random. The default assumption would be that the rates are not influenced by what other people do. In the same way that we assume that it is irrelevant if we throw 100 or 100.000 dice-throws. The expected outcome is that we will have 1/6 of the dices showing 6 in both cases. There will be far more dices showing 6 in the 100.000-lot, but that is irrelevant for the expected rates (which is what we are dealing with rather than absolute numbers).

The studies measures averages squared differences, not total differences which would be absurd, since a larger sample would produce a larger total difference. Demographics may cause the the "clock" to be reset. If there is only one man who fathers all children, then the clock is set to 0 (by his generations). There is however no reason to believe that this has any influence on the mutational process. Even the Zhivotovsky-paper doesn't argue for this.

[Karhunkynsi]

New study about: Ancient watercourses and biogeography of the Sahara explain the peopling of the desert (http://www.pnas.org/content/early/2010/12/23/1012231108.abstract)

Around 11-8kya ago there was several routes through Sahara that could have brought people back and forth. Infact Sahara has been pretty pleasent looking place back then, atleast from anglers point fo view.

[takoja]

I'm thinking, maybe O should be a little more south, closer to Sundaland. Here are few articles.

http://news.bbc.co.uk/2/hi/8406506.stm
http://www.physorg.com/news130761648.html


Dr Oppenheimer’s book, based on multidisciplinary evidence, writes about the effects of the drowning of a huge ancient continent called ‘Sundaland’ (that extended the Asian landmass as far as Borneo and Java). This happened during the period 15,000 to 7,000 years ago following the last Ice Age. He outlines how rising sea levels in three massive pulses caused flooding and the submergence of the Sunda Continent, creating the Java and South China Seas and the thousands of islands that make up Indonesia and the Philippines today.

Martin Richards, the first Professor of Archaeogenetics at Leeds University, who led the interdisciplinary research team, said: ‘I think the study results are going to be a big surprise for many archaeologists and linguists, on whose studies conventional migration theories are based. These population expansions were most likely to have been driven by climate change, in particular global warming and the resulting sea-level rises at the end of the Ice Age between 15,000 to 7,000 years ago.’

Also I've attached a study Male Demography in East Asia: A North-South Contrast in Human population Expansion Times (http://www.genetics.org/cgi/reprint/172/4/2431.pdf) discussin the situation in East Asia before and alter the last glacial maximum.


According to our analysis, the northern populations examined all started to expand in number between 34 and 22 thousand years ago, before the last glacial maximum at 21-18 KYA, while the southern populations all started to expand between 18 and 12 KYA, but then grew faster. We suggest that the northern populations expanded earlier because they could exploid the abundant megafauna of the "Mammoth Steppe", while the southern populations could increase in number only when a warmer and more stable climate led to more plentiful plant resources such as tubers

[Karhunkynsi]

I'm thinking, maybe O should be a little more south, closer to Sundaland. Here are few articles.


Interesting. O's would then be Southeast Asian analogy for I1's and Doggerland (http://en.wikipedia.org/wiki/Doggerland).

[Semitic Duwa]

It is very unlikely that they (or anyone else) were in Middle East. Please take a look at those maps in OP. I doubt there was anyone at Middle East at that time frame (LGM). What does West Asia mean ? Isnt it just another term for Middle East anyway ?


South-West Asia/Middle-East is a fine guess, pilgrim...

If not Syro-Mesopotamia/Taurus-Zagros... Then head for Socotra:)

[Karhunkynsi]

South-West Asia/Middle-East is a fine guess, pilgrim...

If not Syro-Mesopotamia/Taurus-Zagros... Then head for Socotra:)

Cool maps. Based on those I suggest that IJ were just another posse roaming the western half of the Mammoth tundra. I's went west while J's headed south.