I am trying to figure this out myself, but here is how I think it works.
First what does wiring in series mean? It means run a wire from the positive feed on the panel to the positive input side on the device, then from the negative return on the device, run that wire to the positive on the next device & repeat until last device on that zone or circuit, then run the wire back to the negative terminal on the panel - creating a 'daisy chain' (same with the 'alarm wire'. But with modern panels you should try to design your layout to use a separate circuit for each area of the property you are trying to protect. Circuit = Zone. Similarly you can daisy chain the tamper circuit in the same way. On both of these the alarm microprocessor sees a low resistance (for normally closed devices) = closed circuit, no intruders, or high/infinite resistance, when either intruder enters or someone cuts the tamper circuit = open circuit = alarm condition.
The thing is that if these wires bridged at any point then you take the device out of the circuit and the microprocessor is none the wiser = no alarm.
(NB some people use one device per zone, but still have one tamper continuous circuit daisy chained at the panel using a mini 'chocolate block connector' = so called global tamper circuit)
So imagine you have that daisy chain of devices, then 'end of line' means wire the resistor at the last device in the chain before the wire comes back to the panel. But if you only have one device per circuit/zone, then each one is at the 'end of the line', so wire each one up with resistors.
How come resistors? Well the microprocessor can be programmed to recognise not just vey low resistance (=normal operation) and extremely high resistance, but values inbetween these extremes. The advantage is that if someone now tries to bridge out the device, the resistance drops and the microprocessor notes that = alarm condition.
Not sure if I have it exactly correct and there must be some other alternatives for circuits with normally open devices etc., and I am not sure why it has to be last device in the chain, or how that protects the preceeding devices in that chain from bridging.
Another thought is that daisy chaining too many devices on one cicuit/zone may exceed the current capacity of that zone, then possibly the devices might not work correctly or you might fry the panel - not sure of consequences.
ps Read the Texecom manual and and when using resistors 'EOL' method, there is one resistor across each PIR (across input & output for the detector, what they call in parallel with the detector), plus one resistor in the Tamper line. PLUS if I have read the diagram correctly the device alarm sensors and the tampers are on single circuit, i.e. in series. The legs of the resistor can serve this function in one detector, then use a link in the others.
