PDA

View Full Version : Draw-wing help please



drumgerry
25-01-2013, 08:27 PM
1404Hi all

I'm trying to work out how to use Draw-wing and there's a paucity of resources on the web it would seem. I'm a complete novice and have had this sample sitting in the freezer since last summer. I've attached a jpg version of the scan I took at 2400dpi - I have a gigantic bmp file of the same but the version you see here has been darkened a bit to give a greater contrast between light and dark.

When I try to get Draw-wing to attach its points to the wings ("Apis Junctions") it's only picking up one wing from the image instead of the 20 I have scanned. I'm trying to get it right with this scan before I scan enough for a proper sample. I have Morphplot sitting waiting for when I do.

Any suggestions?

Jon
25-01-2013, 09:40 PM
Put a tick in the 'step by step' box
Go to 'file' 'open' and open your scan
Select 'wing' and 'apis junctions dir...'

It should all follow on from there.
You might have to move some points as it misses the odd one.

Peter
25-01-2013, 10:26 PM
Did you scan by transmitted light, i.e. in filmstrip mode. Your image looks like it was scanned by reflective light.

Please read and follow the instructions in MorphPlot precisely.

Jon
25-01-2013, 10:34 PM
1405

That's my setting for the V330 if the scan is the issue

drumgerry
25-01-2013, 11:13 PM
Errr...I don't know Peter. I have a Canon Pixma MP620 and I don't think I have that option. I looked again at the Morphplot instructions and tried again. Used the step by step mode and suddenly realised that if I clicked "next step" another wing image was picked up. Doh!:rolleyes: So in the end Draw-wing picked up 17 of the 20 wings on my scan.

For all the good it did me. It's not looking so good as you'll see from the scattergram on the 2D plot thread. It looks like they're pretty hybridised - which is a shame as I grafted from that queen last summer. Still you have to start somewhere.

drumgerry
25-01-2013, 11:15 PM
Oh one other thing. Although Draw-wing was picking the individual wing images it wasn't placing the points on them. I had to manually do that for all 17 images which took bleedin ages!

Peter
25-01-2013, 11:55 PM
Sounds like a scanning issue. If you are scanning by transmitted light then DrawWing will not work.

You will get quicker at placing/correcting points manually. Again,follow my suggestions in MorphPlot.

Hors
26-01-2013, 10:48 AM
Hi friends of morphometry,
I have here a comparison of two measurement systems.
Wings 1: mother copulated in Virelles / Belgium
Wings 2: daughter (F1) copulated in Hamm / Germany.
Strangely, the high share of mellifera Rigional pairing.
Regards, Horst

1407

1408
PS. work with transparency scanner at 3200 dpi. It is a very good workmanship of the wings in the slide frame, behind glass.

drumgerry
26-01-2013, 01:21 PM
This is the sort of quality I'm getting with my scanner (image of this wing extracted by Draw-wing). Is this ok?

1409

The Drone Ranger
07-04-2013, 01:41 PM
Hi Drumgerry
you probably have this sorted now but just in case
There are loads of reflections on the wing sample scan --that isn't good
Secondly the scanner you are/were using is a CIS type not CCD
CIS scanners typically have a very shallow depth of field which is ok for documents but not for objects
And as already mentioned a transparency scanning mode is likely to be needed
I don't do wing scanning so the last point might be overkill
Here's the Canon spec http://www.canon.co.uk/For_Home/Product_Finder/Multifunctionals/Inkjet/PIXMA_MP620/index.aspx?specs=1

The Drone Ranger
07-04-2013, 01:42 PM
]

This is quite a good page on scanner technologies http://www.imageaccess.de/?page=WhitePapersThePixelRace&lang=en

drumgerry
07-04-2013, 02:27 PM
DR it turned out fine in the end and I was able to get sufficient detail in the scanned image with my scanner - where Draw-wing wasn't able to place the points in the right spots I managed to do it manually with a bit of guidance in the main scanned wings thread. Only had the one sample to do in the winter but with the new season just around the corner I'll be taking a few samples from my own and other people's bees to get a bit more practice in.

The Drone Ranger
07-04-2013, 02:35 PM
That's great news
These things are not blooming cheap
I used to work for Canon at one time and loyally bought an Epson Lol!!
(it works with Linux)
If you ever buy another one a CCD is more forgiving
Sleet here today again

drumgerry
07-04-2013, 02:38 PM
It's a great printer as well so I was reluctant to buy another if the Canon could be made to function which thankfully it was.

Had about 2-3 inches of snow overnight but it has shifted quickly and there's not a lot left on the ground now!

The Drone Ranger
07-04-2013, 02:44 PM
I feel better about the sleet now :)

This chap claims he knows you ??

1484

The Drone Ranger
25-07-2013, 02:49 PM
Surely there is no need for points numbered 1 -19 ?
6 would be boring enough to organise and check
sorry 0 -18 :)

Peter
26-07-2013, 09:28 AM
DrawWing was designed for geographic morphometry (where the shape of the whole wing is measured). It just happens to identify the points that we need for standard morphometry.
We need 0, 1, 2, 3, 4, 6, and 18 (and 5 if you want the Hantel index).

The Drone Ranger
26-07-2013, 09:41 AM
Hi Peter
Thanks a lot, that will save my eyesight
I'll only correct those 7 points in future

The Drone Ranger
20-10-2013, 09:33 PM
Winter is coming there should be dead bees aplenty on the floor
good opportunity to practice your wing plots ??

Peter
21-10-2013, 09:29 AM
Winter is coming there should be dead bees aplenty on the floor
good opportunity to practice your wing plots ??

Not really. You need young bees from the centre of the brood nest so that you can be sure they came from that queen.

Dead ones could be drifters - or robbers from another hive.

The Drone Ranger
21-10-2013, 07:19 PM
point taken Peter but they still make practice subjects I guess

Spring would be the best time before any drifting and when all the bees are sure to belong :)

Peter
22-10-2013, 08:01 PM
point taken Peter but they still make practice subjects I guess

Spring would be the best time before any drifting and when all the bees are sure to belong :)

Hmmm. So what about bees that drifted in the previous autumn. And there is no drifting in the spring? I think not.

Taking them from the centre of the brood nest is the only way to be sure. Give the comb a gentle shake to remove the older bees and then take your sample from those that remain. Why do all the morphometry work if you cannot be sure that the bees came from that queen?

Best wishes

Peter

Jon
22-10-2013, 08:22 PM
Peter.
If you take a sample from the floor and take a sample from the brood nest you will find there is a pretty good correlation.
I have also found that a colony produces a similar scattergram if sampled several times over a period of years irrespective of any seasonal drifting.
For several years now the Bibba magazine has been full of well intentioned articles about the accuracy of sampling techniques, but the elephant in the room is whether there is actually a good correlation between wing pattern and underlying DNA. If not, the arguments about sampling from the floor or from the brood nest are not that relevant.
Cart before the horse.

Wing venation will indicate a hybrid fairly clearly but it may not tell you much about whether the colony tested is pure race especially if you have been sampling wings over a period of years and using that information as part of the selection process. It is quite conceivable that you could get a scattergram with all the points in the correct quadrant which turns out to have mixed genetics.

What does produce very inaccurate results is sampling too soon after a colony has been requeened. Even after 3 months you still have bees from the previous queen present.

The Drone Ranger
22-10-2013, 10:16 PM
Could be wrong but I think winter bees are laid purposefully by the queen
Any foragers or flying bees from this year will be dead by next spring (on the floor right enough)
That's why September on is so important laying wise
Drifting also depends on flow because empty handed bees are unwelcome when the turn up at the wrong door :)

mbc
22-10-2013, 10:26 PM
Drifting also depends on flow because empty handed bees are unwelcome when the turn up at the wrong door :)

I'm not sure how true that is.
Certainly bees bringing forage will mostly be very welcome, but even with nothing to offer I think most healthy bees will get assimilated if they arrive with a non aggressive demeanor.

The Drone Ranger
22-10-2013, 10:45 PM
I'm not sure how true that is.
Certainly bees bringing forage will mostly be very welcome, but even with nothing to offer I think most healthy bees will get assimilated if they arrive with a non aggressive demeanor.
I'll be testing the theory at all the bring a bottle parties this new year :)

Jon
22-10-2013, 10:53 PM
Make sure it's not a bottle of Buckfast (http://uncyclopedia.wikia.com/wiki/Buckfast).

Peter
22-10-2013, 11:13 PM
[QUOTE=Jon;22146]Peter.
>Wing venation will indicate a hybrid fairly clearly but it may not tell you much about whether the colony tested is pure race especially if you have been sampling wings over a period of years and using that information as part of the selection process.

>It is quite conceivable that you could get a scattergram with all the points in the correct quadrant which turns out to have mixed genetics.

Seems to me that the first statement contradicts itself and that is re-inforced by the second.

>the elephant in the room is whether there is actually a good correlation between wing pattern and underlying DNA.

Yes - and this was something that the money from the Co-operative was supposed to resolve, but has so far failed to do - spectacularly!
However, we should perhaps remember that the Galtee group started by using wing morphometry; I seem to recall that those schoolgirls analysed 21,000 wings - and more than earned the top prize for their work.
Of course, none of us are claiming that wing morphometry is sufficient alone, but it served Galtee members well - and it would appear to have worked for me too.

Peter
22-10-2013, 11:14 PM
>Spring would be the best time before any drifting and when all the bees are sure to belong :)

Define spring!

Jon
22-10-2013, 11:25 PM
[QUOTE=Jon;22146]

Seems to me that the first statement contradicts itself and that is re-inforced by the second.

Not at all.
Check out the Moritz paper (http://citeseerx.ist.psu.edu/viewdoc/download;jsessionid=40E944022B5FBCB674B52745189501 63?doi=10.1.1.69.393&rep=rep1&type=pdf)!

Once you start using wing morphometry as part of a bee breeding programme it starts to lose its value.
Moritz recommends discarding it as a selection method or switching to a different set of morphometric variables after a few generations.
This paper is critical to understanding wing morphometry limitations.
The more it is used the less predictive it becomes.
As a snapshot of a population of bees which have never been selected, it should be quite useful.


However, we should perhaps remember that the Galtee group started by using wing morphometry
The wing morphometry survey started about 5 or 6 years ago and the Galtee breeding programme has been going more than 20 years.
DNA microsatellite analysis is getting cheaper every year so this is going to make the wing morphometry redundant in the long run anyway.

The Drone Ranger
23-10-2013, 09:44 AM
Define spring!
Hi Peter
I would say March round these parts but lucky people further South might have them flying earlier