An LED is a diode. As such it has a forward voltage drop, i.e., when it’s conducting current and shining light, there is a small voltage drop from one end to the other. (After all, with nearly zero resistance, its power consumption has to show up somehow—you can’t get something for nothing.)

Different color LEDs have different forward-biased voltage drops. Blue is typically higher than other colors. So when different types of LEDs are wired in parallel, the current will take the path of least resistance (or least power consumption) and the current will not be the same through each device. In your case the orange LEDs are “stealing” all the current from the blue ones.

The solution is to use a separate resistor for each LED in parallel, so the current is forced to divide more or less evenly through each device. Another alternative is to wire them in series, though this won’t work with only 3 volts driving 4 or more LEDs, because the total voltage drop will exceed the battery voltage.

Even if the LEDs were all identical colors and ratings, chances are they wouldn’t conduct exactly the same current through each device & some would shine brighter than others. So it’s still best to give each LED its own resistor.

@gasman Thanks so much! That explanation makes perfect sense to me.

@gasman Actually, a follow-up question. Do I need to use different resistor values for different color LEDs?

In theory if you used exactly the same resistance for each color, there would be less current flowing through the blue than through the orange. I’m not sure, however, if this means it would look less bright (actually I’ve never played with blue LEDs—they weren’t invented yet when I used to do electronics.) I’d experiment a bit, using a breadboard or other simple rig to see what resistances give the best results.

Actually, I would think that each color would require a different resistance value.

The voltage drop across each branch of a parallel circuit will be the same, so if you get resistors with values such that (R1 + Blue) = (R2 + Orange) then things should be good.

This site gives this data for LED forward voltage (VF typ.)
Red 1.7 V
Yellow 2.1 V
Green 2.2 V
Blue 4.5 V

Picture a mixture of these diodes with all their anodes connected together & all the cathodes connected together. As the voltage between these nodes increases to 1.7 V, Red will conduct while all other colors will just sit there, not conducting because they’re below their forward voltage thresholds. Meanwhile the current-limiting resistor that’s supposed to handle the combined currents of all the LEDs is sized too small, so the red LED might fry from over-current.

Also note that a blue LED will never light using a 3V battery. Looks like you need over 4.5 Volts to move current through it.

@jerv It’s tricky to think of LEDs as series resistors, like in your equation. They’re weird—as in highly non-linear. The voltage barely changes as the current increases from minimum to maximum.

@gasman My blue LEDs must be somewhat different, because they do light up when I wire them alone to the battery. Thanks for the rest though! This is good info.

@gasman Maybe for you, but I can’t understand why you wouldn’t think of them that way, so I guess we’re even.