[gakgx6]

I know that increased branching causes a decrease in boiling point for an alkane, but the examples of branching I've seen have always been really obvious. However, in this case both molecule have 2 branching just on different carbons.

My guess is that 1,1 dimethyl butane is considered more branched, therefore it can form a more compacet three-dimensional structure, which packs more easily into a solid structure with a higher melting point and lower boliling point compared with 2,4 dimethylbutane.

Is there any explanation why branches on the same carbon vs branches on different carbon make molecules form a more compact structure? 

[Enthalpy]

Some sources of raw data:
http://www.umsl.edu/~chickosj/JSCPUBS/mp.pdf
http://www.chemicalland21.com/info/Alkane%20Compound%20Boiling%20Points.htm
http://www.maa.org/joma/Volume8/Burch/bpdata.xls

From the data I parsed, branching has a limited effect on boiling point, but a huge one on melting, and this is the key to a wide liquid range.

I believe to understand - until seeing contradictory examples - that branching prevents easy stacking of molecules, and proper stacking with many contact points gives a strong intermolecular bond hence a higher melting point.

In the case of almost linear linear with a few small branches like methyl, the melting point seems to be linked with the length of the free straight tail, hence your 2-methylpentane stays solid at a higher temperature than 2,4-dimethylbutane.

For instance, 2,4,6-trimethyl-tridecane is liquid between -100°C and +250°C approximately, hence I like it as a rocket fuel that won't freeze on Mars and has a huge flash point.

At least one long straight tail also lowers the autoignition temperature, useful for a quick flame (hence heptane burns quickly and iso-octane slowly). I suppose autoignition before boiling further helps quiet burning.

Methyl groups lower the heat of formation; ethyl ramifications would be better for a fuel.