With an incandescent bulb, the voltage rating is the critical factor. Exceed that, and you shorten the life. Exceed it enough and you make a flashbulb out of it. An ideal power supply for a 1.5V light bulb is a rechargeable battery - most rechargable types put out 1.2 volts. Slightly under voltage will make the bulb blow without being a super bright and will extend the life. 14-16V bulbs are good at 12V.
In parallel, voltage is the same to each device in parallel. In parallel, the total current drawn by each device adds up. So for 4 1.5V bulbs of 30ma each in parallel, it's 1.5V at 120ma.
In series, the voltage adds up but the current is the same across the entire string. So those same 4 1.5V 30ma bulbs in series would need 6V at 30ma.
A light bulb is an example of a resistive load. In simple terms, you cna supply that 1.5V 30ma light bulb with 1.5 volts at 200 amps and nothign bad will happen (unless you short out the connecting wires). The lamp will draw 30ma - amps don't 'pump in', the circuit draws what it needs, so you need to supply at least the required, supplying higher only means the pwoer supply can work less hard. ANother example - your house likely has a 200 amp incoming service. When you plug in your vacuum cleaner, it doesn;t get hit with 200 amps and explode, it draws whatever the rating is - probably around 4-5 amps, and works fine.
LEDs work a bit differently. They are current driven, not voltage driven. They operate within a specific current range, exceeding the upper limit will burn them out, just liek excessive voltage burns out a light bulb. There's also a minimum voltage required by LEDs, to overcome the the semiconductor junction that gives off the light., Once this point is reached, teh voltage across an LED is essentially constant. There IS an upper llimit which will break down the semiconductor junction (ie, destory the LED) so you can't just feed 200 volts in, but the point being that is the LED is say 3V, you can run it off a 12V power supply with no problem - the key factor being the current. The LED specifications will list the maximum current it an handle - depending on the LED, 20-30ma is common. SO how do you control the current? The easiest way is with a series resistor. Remember from aobve, in parallel voltages add, but currents are equal. We know a few things: the voltage of the power supply (12V example), the voltage across the LED from its rating (3V example). From this we can calculate the voltage across a resistor in series with the LED: 12(power supply) - 3 (LED) = 9 volts. Wel also know that the LED can stand up to 25ma. But we don;t want to run at the maximum, the LED will be plemty bright at less, plus if the power supply exceeds 12V and we are running at the maxmium, we will exceed the limit. So let's say we want 15ma in the LED, well within the limits. Since currents through devices in series are the same, that means the resistor will also see 15ma.
The next bit if Ohm's Law: Voltage = Resistance X Current. Voltage in Volts, Resistance in Ohms, Current in Amps. 1ma = .001 Amp, so our 15ma number is .015 amps. ANd we know the voltage, 9 volts. Solving for Resistance, we have Resistance = 9 Volts/.015 amps = 600 ohms. Now, reistors have a tolerance range, common cheap ones we use have a 10% tolerance rating, and so don;t come in every possible numeric value. Standard values are listed here: http://www.elexp.com/t_eia.htm So we need to pick somehting close to the calsulated value. It's generally safest to pick the next HIGHEST standard value, in this case 680 ohms. Plugging that 680 back in and calculating the current, you have Current = Voltage/Resistance = 9V/680 ohms = .013ma, still definitely a safe value.
Clear as mud now? The real key to understanding any of this stuff is the behavior of circuit elements in series and parallel with respect to voltage and current, these are derived from Kirchoff's Laws, and Ohm's Law, that's the Voltage = Resistance x Current equation. Please don;t look these up on something like Wikipedia unless you understand calculus, as they are expressed in calculus terms which is their true root meaning but it really all simplifies down to the simple terms I tried to state it in.
And a note to other EE's - I am an EE too, and I understand that in most cases I am oversimplyfying and assuming ideal circuits which don't really exist in nature., However, for the purposes of the use here, in model railroading and simple hobby electronics, it's more than accurate enough. Plus if these methods didn;t work, every headlight in every one of my locos would have blown up a long time ago