Why do some multimeters have dedicated mV ranges?

[Robomeds]

I was looking at a Fluke 73iii I own and started thinking, why does it have a dedicated 300mV range?  I can understand why someone might want to measure a low voltage.  I'm not questioning the desire to measure sub 300mV ranges but really, why can't/isn't this incorporated into the autoranging DC voltage range?  Is there a significant change in input impedance in the mV range?
Incidentally I have seen this on several of my non-Fluke meters as well. 

[Monkeh]

Is there a significant change in input impedance in the mV range?

Only about 2.99GOhm, typically.

[Robomeds]

gigaohms.... hmmmm.... 

[Monkeh]

Your Fluke seems to be 10M on mV, though. Pays to check the specs..

Not sure why it has a separate range for the same impedance.

[Robomeds]

Well the 87-iii and 73-iii show the same impedance in both the mV and V settings according to Fluke.  However, I do see that if I start the meter in high impedance mode the 400mV range does gain 3.99 Gohms of impedance.  By default the 87-iii has only 10Mohm resistance in either DC volts range.  That could explain why the 87 has such a range but not the 73-iii which doesn't have a high impedance setting. 

[ModemHead]

On Fluke auto-ranging DMMs, there is usually a 10Meg resistor in series with the input, and one of several other grounded precision resistors is switched in by the microprocessor to form a divider for 1:10, 1:100, etc.  But for the millivolt range, the 10Meg resistor is grounded, and the ADC input is taken straight off the input jack (1:1) through the ohms sense resistor or through a separate resistor of about 1Meg.  The input impedance thus remains at about 10Meg for all ranges.

Given this input divider reconfiguration, I suppose it's easier to just have another switch position.