The three stages of the Bel Canto technique are as follows:

The "lift of the throat:"  

In Bel Canto, the voice is NEVER directed towards the soft palate or the throat. This makes sense.... consider: you can whack a bean bag chair with a Louisville slugger and not make much noise... hitting a soft surface (ie soft palate, throat) is useless. You're much better off tapping your knuckles on the lid of a garbage bin. (less effort, harder surface, resonating chamber....) A lot of singers think they have to yell and scream to get any projection, but those voices tend to be absorbed in part by the soft palate. A lot of effort for little gain. Directing the voice at a hard surface (teeth, hard palate...) will give much better projection. 

Try smiling when you sing. Not a cheezy grin like someone is driving a nail into your baby toe kind of smile, but a more "open and relaxed" smile - like someone just complemented you on your new shirt, or like you've just recognized a friend in the crowd. By visualizing separating the upper and lower jaw at the back teeth, visualizing yawning in the back of the throat to open the airway, and singing with a relaxed smile - mouth turned up gently at the corners - the throat is relaxed.  If you actually DO yawn, you're probably doing it right. The result of this will be a lifting of the uvula (the punching-bag-thingy in the back of your throat). Your whole throat will be open and relaxed. Do this before you sing a note, and maintain this position at all times when you sing.

The "mask of the face:" 

This is pretty widely used in many techniques. Essentially, the mask is an inverted triangle where the higher notes are thought of as wide and just below the cheekbones, but in reality are placed wide across the dome of the hard palate. (the dome of the hard palate is just below the cheekbones...) Essentially, you are placing the voice "high" on the face.  The lower notes are in the bottom point of the inverted triangle, which puts them just in front of the lips and very narrow.  Focusing low notes very narrow and in front of the face brings in resonance from the chest more than it does for middle and high notes, but regardless of the pitch, the resonance is actually created in the sinus cavities.  In Bel Canto, just behind the teeth is the placement for "mid-range" notes.    This line of thinking is taught in other techniques as well...

The "inhalation of the voice:"

Okay.... here is where everyone thinks you're lost your mind.  They'll demonstrate trying to sing by sucking air backwards through the vocal cords as a means toward showing how silly and impossible it is.  Here is the scoop in a nutshell. 

The sound is actually drawn in from outside the body, (involves visualizing a "pulling in of the breath" ) with a stream of air hitting the hard palate, and passing over (not through) the vocal cords, and into the sinus cavities. As the air passes over the vocal cords, it creates a difference in air pressure, much like when a truck passes by and pulls everything along with it, and draws air up from the lungs as it passes over. The air that is drawn up from the lungs passes through the vocal cords the right way, creating the sound. The inhaled breath bouncing off of the hard palate creates projection, and the air going into the sinus cavities provides resonance. Lamperti wrote (quoting roughly), "sing as if continuing to take in breath."  The difference is that the air is not pushed up and out from the lungs with the diaphragm muscles (though the diaphragm is important), but rather drawn out of the lungs by air passing over the vocal cords. 

This results in two upward streams of air (one from the mouth heading up into the sinus cavities and one up from the lungs and through the vocal cords). These two streams of air meet in the sinus cavities. When two rivers meet, you will get a whirlpool at the point of intersection. The same happens when two streams of air meet. In the case of Bel Canto singing, this "whirlpool of air" occurs in the sinus cavities, resulting in a rich resonance.  The air passes from the sinuses through the nose, though because the tone has already been achieved, there is no nasal quality to the tone. 

Because air is not being forcefully pushed out of the lungs, there is no strain on the vocal cords (providing you're doing a number of other things right, like lifting the throat, etc.), and therefore no fatigue of the voice. 

It is important to know where and how to inhale the voice. You CAN NOT inhale the voice into the lungs, and expect the vocal cords to vibrate properly as the air passes backwards through them. That would be contrary to human physiology.

The "hold of the breath:"