Sea water is a good electrolyte, a boat is full of dissimilar metals bolted together, and the result is a battery you did not draw. What decides whether it matters is not mainly which metals — it is the area ratio between them, and whether there is a continuous wet path from one to the other.
Three things have to be true at once
A galvanic cell needs all three, and removing any one of them stops it:
1. Two metals at different potentials. The separation in the practical galvanic series is the driving voltage. Below about 0.15 V it is usually not worth designing around; above 0.25 V it is; above 0.5 V it will find you.
2. An electrical connection. Through the fastener, the bracket, the hull, the bonding strap — anything metallic. An isolated bush or washer breaks it, but only if it breaks every path, and the one you forgot is the bilge water.
3. An electrolyte bridging both. Sea water, or a damp salt crust, or a wet gasket that never dries. This is the one you can actually design out: keep the joint dry, keep the crevice bedded, and the couple is academic.
Area ratio is the part people get wrong
The corrosion current is set by the cathode area, because that is where the reduction reaction happens. That current then concentrates into whatever anode area there is. So the penetration rate on the anode scales roughly with the ratio of cathode to anode area — and that is geometry, not chemistry.
The practical rule follows: a small anode against a large cathode is the arrangement that destroys parts. A stainless bolt in an aluminium plate is survivable — the bolt is the cathode and it is tiny. An aluminium bolt in a stainless plate is the same two metals the other way round, and it will eat the bolt. Same series, same driving voltage, opposite outcome.
Drag the area ratio below and watch what it does to the relative attack rate. Nothing about the metals changed.
Try it — the same two metals, different areas
What to do about it, in order
Choose the couple. The cheapest fix is the one made on the drawing. In sea water the safe neighbours for aluminium are other aluminium alloys and the 5xxx family in particular; stainless, bronze and graphite are all a long way noble of it. Carbon-fibre laminate is the one that surprises people — graphite sits at the noble end of the series, so a carbon panel bolted to an aluminium frame is a severe couple with an enormous cathode.
Get the ratio the right way round. If you must mix, make the noble metal the small part: stainless fasteners into aluminium, not the reverse. And if the fastener is the anode, it is the part you will be replacing.
Break the electrolyte path. Bedding compound, a sealed joint, drainage so salt water does not sit in the crevice. This is worth more than most coating schemes, because a crevice that stays wet also goes anaerobic and starts crevice corrosion on its own account — which attacks passive stainless where galvanic action alone would not.
Insulate, carefully. Isolation bushes and washers work, but a coating on the anode is worse than no coating: every pinhole becomes a tiny anode against the whole cathode, which is the bad area ratio at its most extreme. Coat the cathode if you coat either one — that shrinks the driving area and the current with it.
Then protect cathodically. An anode is what you fit when the couple is unavoidable and wet: it makes the whole assembly the cathode of a bigger cell. Size it properly — the anode sizing tool does the DNV-RP-B401 arithmetic, and the number it gives is usually more metal than people expect.
Aluminium plate with stainless fasteners, in sea water. Which one do you paint?
Coating the anode is the classic own goal: the paint is never perfect, and every pinhole concentrates the whole cathode’s current into a spot of bare aluminium. Coating the cathode reduces the area driving the reaction, so the current falls and what is left is spread over the full anode area. Coat both if you like — but if you only coat one, coat the noble one.
Aluminium hulls have an upper limit too
Steel wants to sit at or below −0.80 V and more negative is simply safer. Aluminium does not work that way: drive it much past about −1.1 V and the cathodic reaction makes hydroxide at the surface faster than the passive film can cope, and the metal you were protecting starts to dissolve in its own alkali. That is why magnesium anodes do not go on an aluminium hull in sea water — magnesium's driving voltage is far too high. Aluminium anodes on aluminium, zinc if the water is brackish, and a reference-electrode reading rather than a guess if it matters.
The other asymmetry worth knowing: stray DC current from a badly wired shore supply or a bonding fault will out-corrode any galvanic couple you can design, and it does it in days rather than seasons. If a boat is losing anodes in weeks, look for a wiring fault before you re-specify the metals.
What the series does not tell you
The practical galvanic series is measured in flowing sea water and it is a ranking, not a rate. It says which metal will be attacked and roughly how hard the cell is driven; it says nothing about how fast, because that depends on area, temperature, flow, oxygen and how well each surface passivates. Stainless in particular appears twice — passive near the noble end, active near mild steel — and which one applies depends on whether oxygen can reach it. In a tight, stagnant crevice it cannot, and a fastener that was noble on the bench becomes anodic in the joint.
Use the galvanic series and the area rule for the ranking, and treat the rest as a reason to keep sea water out of the joint in the first place.